Slashdot Mirror
Jordan Pollack Answers AI And IP Questions
It's still just 10 years or so away.
It's not getting any closer.
How do you justify any degree of optimism about the future of AI at this point? What makes now fundamentally different from anytime in the past 40 years?
It is funny, this is the same question I asked Marvin Minsky, the father of AI, at ALife 5 in Japan. He attacked every modern approach, including neural nets, fuzzy logic, evolutionary algorithms, and so on for over an hour, suggesting that his student's (Winston's) thesis should have been the paradigm of the field! I asked, "If AI sucks so much, why are you still in the field after 40 years?"
Hypocrite! Here I am, still in the field after 20 years! As soon as I've convinced myself one approach to AI is too slow, I find another, leaving quietly without attacking the friends I've made. AI is a big wide open field with a lot of smart people trying different things. (Savage attacks by insiders exiting are the worst thing in science, such as Bar Hillel's attack on Machine Translation in the 60's. Forty Years later, MT is "cool" again, in this month's issue of Wired.)
So I can say that, from my perspective as having worked on many different approaches to AI, writing problem space search algorithms for solving puzzles will not result in a general problem solver. Automating predicate logic won't make a computer equivalent to a philosopher. A computer can't do natural language any better than Eliza, without an internal need to communicate to survive and a large blessing of custom hardware. Neural nets are great function approximators with good mathematical results on limited kinds of learning, but we can't set 12 weights to get what we want, let alone 10 billion weights. And even though simple nonlinear systems give off chaos and fractals, Kolmogorov's law tells us simple systems are still simple. Evolution is one path to complexity, but most genetic algorithms simply search a finite search space and optimize a fixed goal.
So I'm locally pessimistic but globally optimistic! Who said AI is 10 years away? It's here now, in limited forms, yielding a lot of economic value, as your mouse clickstream is datamined so the ads which pop up are for things you might actually buy. But the SF ideal of a humanoid robot like Commander Data is centuries away.
I hold the view that any system which responds to its environment in a conditional way based on some internal state, even a thermostat, has a bit of intelligence. Immune systems, ecologies, and economies design things and solve problems. Every computer program you write has a bit of intelligence captured in it. The problem is, real AI of the sort you are alluding to is an organization which might be realizable as a 10 billion line program or a 10 billion weight dynamical neural system, and no human software engineering team can write autonomous code which is more than 10-100 million lines. Even Windows is just DOS with wallpaper, and big applications always require a human in the loop, selecting subprograms from menus or command lines.
Since 1994, we've been working on how to automatically evolve physical symbol systems which would have 10 billion unique moving parts, what we call "Biologically Complex" systems. When I say "We," it is because everything I do is in collaboration with my Ph.D students! A 10 Billion Line program is an absurd goal obviously, but it drives our research to focus in on the process of growth itself, rather than on what shortcuts we can accomplish by hand. We look at co-evolution, which involves machine learners training each other, and on questions of what kinds of substrates for computing could provide a universe of functionality while being constrained in a way which reduces the size or dimensionality of the search space. This constraint is called inductive bias. We seek minimal inductive bias systems, in which the human hints, or "gradient engineering" tricks are fully explicit. (Sevan Ficici, Richard Watson) We still work on neural nets and fractals as a substrate, and have made some progress in understanding how they work (Ofer Melnik, Simon Levy).
It's been more than five years, and while we are not even at the million line mark yet, I am still optimistic and haven't given up on co-evolution to move to a new field. I think that my lab has made progress in understanding why Hillis's sorting networks and Tesauro's Backgammon player were such breakthroughs and where they were limited. (Hugue Juille, Alan Blair). I think we have begun to understand the nature of mediocrity as an attractor in educational systems and how to change the utility functions to avoid collusion, and apply this to human learning (Elizabeth Sklar). We have become more applied, bring co-evolution to the Internet and to robotics, replicating and extending the beautiful results of Karl Sims from 1994 (Pablo Funes, Greg Hornby, Hod Lipson). All the work is available to study at the laboratory's Web site.
AI and ethics. (Score:5, Interesting) by kwsNI What do you say to the people that feel it is unethical to try to create "intelligence"?
I take this as a shorter version of the longer religious question the editor thankfully didn't select. I've talked to myrabbi, perhaps one of the great theologians around today. Even though I am an atheist, he thinks I am on a spiritual quest to understand [God as] the principles of the universe which allow self-organization of life as a chemical process far from equilibrium which dissipates energy and creates structure that exploits emergent properties of physics. Can a spiritual quest be unethical? I suggest that people with this question read Three Scientists and Their God, by Robert Wright, or watch the Morris documentary "Fast, Cheap and Out of Control".
A second ethical question, besides usurping God's rights, is how can you take funding from national and military agencies like NSF, Darpa and ONR? For the past 50 years at least, they have been the seed capital for the science behind most of the technological progress I know about. With the venture capital economy, that curiosity-based seed function may be privatized, if some of the big VC funds dedicate 10% for long range science, and the ethical question of whether you are doing something for public good or private gain begins to dominate over the religious and military questions. That is the same question many scientists and Linux hackers ask themselves daily: Can I do good and make money without a conflict of interest?
Turing award. (Score:5, Funny) by V. Do we win something if we can fool him into answering a computer-generated question? ;)
It has always been the case that limiting the range of dialog leads to more successful masquerading. In our CEL online educational game, for example, the only interactions between players are the actual plays, which enables artificial agents to be accepted as game partners.
BTW, the Turing Award is an annual lifetime achievement award in computer science, which has gone to people like John Backus for his eloquent apology for Fortran when he should have given us APL and LISP. The Turing Test is the name given to Alan Turing's proposal for testing for successful AI. Given that we don't deny airplanes fly, I think if AI ever flies, we won't question it. So I propose using the Louis Armstrong Test, his answer to the question "What is jazz?"
How should an amateur get started working on AI? (Score:5,Interesting) by Henry House It seems to me that a significant problem holding back the development of AI is that few non-professionals grok AI well enough to offer any contribution to the AI and open-source communities. What do you suggest that I, as a person interested in both AI and open source, do about this? What are the professionals in the AI field doing about this?
Reading is fundamental.
Frankenstein (Score:5, Interesting) by Borealis For a long time there has been a fear of a Frankenstein being incarnated with AI. Movies like The Matrix and the recent essay by Bill Joy both express worries that AI (in the form of self replicating robots with some AI agenda) can possibly overcome us if we are not careful. Personally I have always considered the idea rather outlandish, but I'm wondering what an actual expert thinks about the idea.
Do you believe that there is any foundation for worry? If so, what areas should we concentrate on to be sure to avoid any problems? If not, what are the limiting factors that prevent an "evil" AI?
AI doesn't kill People. AI might make guns smart enough to sense the weight or handsize of the user, preventing children from killing each other. Everything ever invented is capable of good or evil. Evil arises most often when masses of humans are denied fundamental rights. The Evil Rate and Unemployment Rate are closely linked.
I read Bill Joy's article in Wired last month. And I loved the Unabomber's excerpt because it is based on some of the best Philip Dick paranoid Science Fiction, like: Vulcan's Hammer, We Can Build You, and the Simulacrum. There is a lot of SF on the Golem question and one of my favorites is Marge Piercy's He, She, and It , which proposes a moratorium on AI inside humanoid robots. You can have smart software on the Web, and human looking idiobots, but you can't put real AI inside human looking robots, or you have to pay the price.
My lab is indeed working on self-replicating robotics and were worried for a split second about getting the fetal brain tissue reaction when our paper comes out shortly. We can now envision the "third bootstrap", after precision manufacturing and computation, where machines make the machines which make themselves, just as machine tools are used to make more machine tools, and computers compile their own programs. But the replication loop is quite a sophisticated automatic manufacturing process, which requires a large industrial infrastructure, and a lot of liability insurance. So far, no VC's, Saudi Princes, or government agencies have offered the necessary $500M first round of financing for fullyautomateddesign.com.
It would be wrong of me to say leave my frankenbots alone, and go after frankenfoods and frankenano. I think Joe Weizenbaum's book should be required reading, because every few years somebody else comes up with the idea of inserting computers inside animal bodies, so that the first act of any war will be to exterminate all nonhuman life forms. But I do think we have to worry more about large scale industrial and agricultural processes which are allowed to externalize their by-products affecting the environment, than we need worry about robotic ice-9. We will die quicker from e-mail spam caused by viral marketing customer acquisition schemes or from global warming and ozone depletion triggering major climactic change, red tide or another pollutant taking out fish from the food chain, or even from people throwing away old EGA screens and 386 motherboards in landfills, poisoning the aquifers. I promise that for every robot we build, there will be another robot to recycle it when its job is complete.
Anyhow, IMO Joy's angst must reflect the Sun setting on any instruction set architecture besides x86, but that's a different discussion. Talk to me about the ethics, when your very own open source movement leads to the inevitability of an Intel instruction set monopoly by providing a useful alternative to Microsoft :)
Questions based on your academic path (Score:5, Interesting) by Anonymous Coward The way to the field of AI isn't always extremely clear. What type of background do they expect? Is it mostly a researching position or is it treated like a normal job with normal goals? Are there any classes or subjects or schools you recommend to make it into the AI field? Also, how exactly did you get into the field? How did AI intrigue you into what you do now, despite all the controversy to create an intelligence that could possibly be considered a "god" compared to the human existence? Very interesting to say the least, and something I'm interested in.
There is no AI business field to speak of which is differentiated from the general software business. Most companies which were "AI companies" in an earlier generation of university spin-offs for Lisp Machines, and Expert Systems Shells, failed miserably. Venture Capitalists won't fall into the same sinkhole twice. There are industrial process control companies which use refined bits of AI, e.g. in visual inspection of manufacturing processes, and Neural Network companies, like HNC, who have changed business plans and are now "pattern-recognition e-commerce security." companies. The Speech recognition industry has condensed into one company. Web- based AI means search engines and Language Engines. Ask Jeeves and Google and Direct Hit and many others may use bits of AI and adaptive technologies in their system.
Jobs in AI are just like software jobs everywhere: chain you to a workstation and make you work out boring details in exchange for salary and very little equity. But find a great graduate program in computer science, and you will likely find fun and exciting work for no salary and no equity! And you have to be great at both real and discrete mathematics as well as a natural born programming genius.
As for me, I started programming computers in APL as a freshman in college, and because it was such a high level language and I didn't sleep much, I wrote an awful lot of code in a few years. I was naturally drawn to building heuristic puzzle solvers, game players, and logical theorem provers. Before I met my wife, friends thought I was in love with computers. After working at IBM, I went to graduate school in Urbana and worked with David Waltz on LISP hacking, natural language processing, and reinvented neural networks, which were censored from the AI curriculum of the early 80's. I came to the limit of what could be done with neural networks for intelligence by 1988, and at Ohio State University, started looking at fractals and chaos as a source for generativity. Unfortunately, interesting behavior requires lots of levels and lots of parameters, which is why we started looking at evolution for selecting and adjusting lots of parameters, a focus since I've been at Brandeis.
While there is a lot of detailed work and dead ends, the search for
mechanical intelligence is one of the great unsolved problems, which is
in some way deeply equivalent to questions on the origin of life, human
language, morphogenesis, child development, and human cultural and economic
change. John
Casti's book is a great place to start reading about these big problems.
Also, what do you think about the potential of the models used today to attempt to achieve a working AI? Do you think that the models themselves (e.g. the neural net model) are correct and have the potential to produce an AI given enough power and configuration, or do you think that our current models are just a stepping stone along the way to a better model which is required for success?
Obviously there are clear medicinal benefits to brain research. And the study of any real biological system leads to interesting metaphors which can be the basis for a novel computational model. But I think it is unlikely that research into the biology of the brain is crucial to understanding cognition or replicating intelligence. It's like studying the width of wires in integrated circuits of a computer. Even if you get the whole wiring diagram for a computer, it still tells you little about the programs running on it. I think understanding the brain is a problem which is underestimated. I heard 25,000 scientists attend the annual Neurosciences meeting, three times the largest ever interested in AI. It could be called the Mandelsciences meeting, and different labs compete to describe what they find in those little windows on the Mandelbrot set! But I have a lot of friends who are neuroscientists, and I can be just as facetious about linguistics.
Seriously, I believe we have to understand and replicate the processes which lead to the development of the brain and its behavior, not replicate the mammalian brain itself.
The second part of your question "how intelligence has emerged in us" can be interpreted as a more interesting direction. Here, there is a lot of opportunity to relate human intelligence as animal intelligence plus a little more. The fields of evolutionary epistemology, adaptive behavior, and computational neuroethology are quite interesting. It is a great question to understand cognition as it appears in other animals, insects, worms, and even bacterial colonies. The basic principles of multicellular cooperation are more important than the millions of specific adaptations of the human brain.
As for models question, it is sort of like asking whether a chair is built out of metal, wood, plastic, rubber, or cardboard. It doesn't matter, as long as it are strong enough. The organization of molecules has to provide a surface and a normal force at the right height for sitting. As for the organization of 10 billion things which might make an AI? Doesn't matter if it is c, java, lisp, neurons, or tightly coupled markovian 2nd order polynomial fuzzy sets. Will it stand, or collapse under its own weight?
most likely path? (Score:5, Interesting) by Anonymous Coward Dr Jordan:
Do you think that AI is more likely to arise as the result of explicit efforts to create an intelligent system by programmers, or by evolution of artificial life entities? Or on the third hand, do you think efforts like Cog (training the machine like a child, with a long, human aided learning process) will be the first to create a thinking machine?
We are taking the second path, seeking the principles for self-organization so we can harness them to create and invent forms of organization.. There is a 4th path you don't mention, which is the terminator/Truenames hypothesis, that AI will simply arise among the powerful router machines of the internet. How would we recognize coherent behavior arising in telecom infrastructure if it didn't wake up talking English? I think a SETI for coherent intentional behavior emerging out of the infrastructure would be a fun project to do for the people worrying about risks to the information infrastructure.
Software Market & Open Source (Score:5, Insightful) by Breace In your 'hyperbook' about your idea of a software market I noticed that you say that Open Source evangelists should support your movement because it will be (quote) A way for your next team to be rewarded for their creative work if it turns into Sendmail, Apache, or Linux.
I assume (from reading other parts) that you are talking about a monetary reward. My question is (and this is not meant as a flame by any means), do you really think that that's what the Open Source community is after, after all? Do you think that people like Torvalds or RMS are unhappy for not being rewarded enough?
If the OS community doesn't care about monetary rewards, is there
an other benefit in having your proposed Software Market?
According to economic theory, utility is what motivates you to make decisions in your own self interest. Simple games, like the prisoner's dilemma, rationalize utility with numeric values to illustrate the concept, but it isn't money at all. If someone behaves in an unpredictable way, we must have our definition of their utility wrong.
There are plenty of motivations for writing open source code, including the challenge and the feeling of altruism, both of which have utility. A lot of people may write open source for credit in the community, which also has utility. If RMS was a radical advocate of anonymity who wrote the GPL so you couldn't put your name on the source code because it promoted the glorification of the individual, participating might provide less utility.
Why not Write a Screensaver? (Score:5, Interesting) by peteshaw First of all, it is indeed an honor to pester a big name scientist with my puny little questions! Hopefully I will not arouse angst with the simplicity of my perceptions. Aha! I toss my Wheaties on Mount Olympus and hope to see golden flakes drift down from the sky!
I have always thought that distributed computing naturally lends itself to large scale AI problems, specifically your Neural Networks and Dynamical Systems work. I am thinking specifically of the SETI@home project, and the distributed.net projects. Have you thought about, or to your knowledge has anyone thought about harnessing the power of collective geekdom for sort of a brute force approach to neural networks. I don't know how NN normally work, but it seems that you could write a very small, lightweight client, and embed it into a screen saver a'la SETI@home. This SS would really be really a simple client 'node'. You could then add some cute graphics like a picture of a human brain and some brightly colored synapses or what have you.
Once the /.ers got their hands on such a geek toy I have no doubt you'd have the equivalent of several hundred thousand hours or more of free computer time, and who knows, maybe we could all make a brain together! I would love to think of my computer as a small cog in some vast neural network, or at least I would until Arnold Schwarzenegger got sent back in time to kill my mom. Whaddayathink, Jordan? Is this a good idea, or am I an idiot?
No, its very imaginative. You could be one of my AI grad students. But rather than focusing on neural networks, which, because of matrix multiplication, do not distribute well, people are looking at such systems for evolutionary computation. You can evolve individuals on networked workstations and collect them, or evolve populations which interact occasionally and pass dna around. Look at Tom Ray's Net Tierra project to see how it is going. My colleague Hod Lipson is developing a screensaver for our evolutionary robotics project, but release 1 will be Windows rather than Linux compatible (./sorry)
Actually, one of my early business plans for the Internet, circa the first working java browsers, was to show naughty pictures while harvesting cycles from your computer and reselling them to people needing computer time. All was needed was an assembly language interpreter in java and some interfacing. The problem is that most computationally intense problems people want to solve have large data flow requirements which conflict with the download of the naughty pictures! When I recently tried to corner the market in pig latin domain names for my new "incubator", panies.com panies.com, I didn't secure putation.com because it sounded bad. One week later I realized it was a pretty good name for a distributed computation service, but somebody else had grabbed the URL!
However, there is a critical piece missing from all these visions. intelligence is a property of an organization of computation, it is not computation itself. The problem of robotics is not the limited power of microcomputers, since we could drive any robot from a supercomputer if we knew what to write! We can get infinite cycles already, but nobody can write a coherent program bigger than 10M lines. We have figure out to use cycles towards discovery of a process of self-organization, rather than on a known software organization itself.
AI Metrics (Score:5, Interesting) by john_many_jars I have read several coffee table science books on the subject and often find myself asking for a way to measure AI. As has been noted, AI is always elusive and is just around the corner. My question is how do you gauge how far AI has come and what is AI?
For instance, what's the difference between your TRON demonstration and a highly advanced system of solving a (very specific) non-linear differential equation to find relative and (hopefully absolute) extrema in the wildly complicated space of TRON strategies? Or, is that the definition of intelligence?
This is a very hard question which I won't be able to joke my way out of. I think that system performance in specific domains can be measured, like a rating system for a game likeTRON. I think we might be able to get a measure of the generative capacity of a system in all possible environments, by capturing strings of symbols representing different actions, and looking at the grammar of behavior. In general, however, observers have an effect on their observations of computational capacity. I usually think of intelligence as a measurement, not the thing being measured, sort of like the difference between temperature and heat, or weight and mass. It could be a measurement of operational knowledge (programmed, not static in a database), or of efficient use of knowledge resources. This measurement is applied to an organization. So committees of very smart people can operate idiotically, and groups of dumb insects can be very intelligent.
My current best working definition is that intelligence is the ratio of the amount of problem-solving accomplished to the number of cycles wasted. When I say we need 10B lines of code, it is not to say that raw program size is a measure of intelligence, but to express the idea that inside that code are enough different heuristics and gizmos to solve lots of problems effectively.
And what about Freedom? (Score:5, Insightful) by Hobbex Mr. Pollack,
I read your article about "information property" and was surprised to find you dealt with the matter completely from the point of view of advancing the market. Their are those of us who would argue that the wellbeing of the market is, at most, a second order concern, and that the important issues that Information age gives rise regarding the perceived ownership of information are really about Freedom and integrity.
These issues range from the simple desire to have the right to do whatever one wants with data that one has access to, to the simple futility and danger of trying to limit to paying individuals something that by nature, mathematics, and now technology is Free. They concern the fact that our machines are now so integral in our lives that they have become a part of our identity, with our computers as the extension of ourselves into "cyberspace", and that any proposal which aims to keep the total right to control over everything in the computer away from the user is thus an invasion into our integrity, personality, and freedom.
Do you consider the economics of the market to be a greater concern than individual freedom?
This is a beautiful question, thank you. My book is exactly about freedom and rights: The freedom to sell a copy of a book you are done reading. The freedom to share in the rewards when something you design or write is in demand by millions of people. The right to own what you buy.
I see an inexorable movement towards dispossessionism, both coming from the "right," with UCITA, secured digital rights, anti-crypto-tampering in the DMCA, and ASP subscription models, and coming from the "left", with ideas that we should give our writing up into free collectivist projects.
The Internet is the beginning of Goldstein's "celestial jukebox," the encyclopedia of everything anyone has ever written, every episode of every TV show, and every song by every band. It sounds wonderful until you realize that you will have to pay per view! Bill Gates now has the money to deploy satellites which will force you to rent his word processor for $1/hour, the same rate for renting a movie. The laws on theft of satellite programs, unfortunately, as legal doctrine goes, considers decoding satellite broadcasts as theft of cable services, rather than as protected first amendment rights to receive radio broadcasts. Once secure distribution of programs on a rental basis is established, all content publishing will move inexorably into that mode to maximize profits. No more books, no more records. No more ownership. Dispossession.
The Free software movement, League for Programming Freedom, Open Source Software, on the other hand, talk idealistic young individuals out of their writing. "Contribute it towards a greater good." Be rewarded by occasional e-mails of thanks from your peers. The Free Music movement, or "let's RIP our CD's and trade MP3s through Napster" isn't as politically as economically motivated, but is also making musicians contribute their work for the greater good, at least of dormitories! Dispossession.
Fascism and Communism, while they have philosophical appeal for their mimetic simplicity, have proven themselves consistently the enemies of freedom, enterprise and creativity. Ordinary people are "dispossessed" of their property, which ends up, not surprisingly, in the pockets of the promoters of the simple philosophy.
My purpose in writing License to Bill is to begin a discussion not only on a societal remedy to the microsoft problem, but to secure, as a human right, the right to own information properties I buy, rather than just being able to rent them. I especially want the right to own and sell copies of my own creations, and to own a library of other's creations, reasonably priced based on supply and demand, without fear that a change in technology will render my investments worthless..
A market is just a mechanism which humanity uses to allocate resources fairly. It is neither good nor evil.
To which I would add... (Score:5, Interesting) by joss I also read your IP proposal, and agree with the points mentioned above.
However, I also have a problem with your proposal from an economic perspective:
Property laws developed as a mechanism for optimal utilization of scarce resources. The laws and ethics for standard property make little sense when the cost of replication is $0. The market is the best mechanism for distributing scarce resources, so you propose we make all IP resources scarce so that IP behaves like other commodities and all the laws of the market apply.
We are rapidly entering a world where most wealth is held as a form of IP. Free replication of IP increases the net wealth of the planet. If everybody on earth had access to all the IP on earth, then everybody would be far richer - it's not a zero sum game. Of course, we're several decades at least from this being a viable option since we've reached a local minima. (Need equivalent to starship replicators first - nanotech...)
Artificially pretending that IP is a scarce resource will keep the lawyers, accountants, politicians in work, and will also allow some money to flow back to the creatives, but at the cost of impoverishing humanity.
I could actually see your proposal being adopted, and I can see how it will maintain capitalism as the dominant model, but I also believe that it is the most damaging economic suggestion in human history
Could you tell me why I'm wrong.
Wow! "I also believe that it is the most damaging economic suggestion in human history" Surely this is a wonderful compliment.
The history and future of money is very interesting, and one you can read about in various books, including one byMilton Friedman, and one from the Cato Institute. I think today's software houses who force upgrades on their customers are like the wildcat banks of the nineteenth century, printing up banknotes, and then declaring bankruptcy, vanishing with the deposits and setting up shop in another town.
Before money, there was simply trade in raw and polished goods. Then there was weighing and coinage. Lots of people thought coins were the real value and heartily resisted paper money. The gold and silver standards gave way, and eventually the idea that there was gold for every dollar bill was revealed as a hoax, and now "money" is simply a record in your bank's computer that there is a certain amount you are entitled to withdraw based on the amounts other banks have deposited for you. The only essential different between a rich and poor person is what the bank computers and the registrar of deeds say it is, backed by military force. And the money supply and international exchanges now somehow represents our national wealth with respect to other nations, and other nation's confidence that our banking system isn't duplicating dollars. Instead of objects of trade, money is information about potential trade.
While you might not like the idea that money is abstract and in limited supply, and you have more or less than you want, it is the soft underbelly of "Starship Economics" that Gene Roddenberry died before coming up with the backstory for how to have a non-mediocre society with unlimited replication for all.
I once invented a transporter machine for paper using public key crypto and fax technology. It would hold the source paper in a metal box, verify the copy was printed, and then destroy the original and legitimize the copy. With this system, you could fax a dollar bill to a friend! Now: is a dollar bill is just the likeness of a dollar bill on a crinkly piece of thermal paper, or the actual piece of green stuff? If Paypal can figure out how you can beam money from your palmpilot to mine, but a bug lets you keep a copy of the money, I bet their valuation would go way down.
I am simply saying that permanent use and resale licenses to changeable information (software, art, literature, music, movies) which can be traded securely, without loss or duplication, in a public market, is a form of currency.
Unlimited replication of currency just doesn't work, any more than two copies of William Shatner.
I stake the middle ground. Both the "right" copyright publishers who make currency loss through expiring keys and forced upgrades, and the "left" copyright violators who duplicate currency, will be welcome at my table when they see the light.
----------
Thanks for your interesting questions. My comments do not reflect the official position of my employer Brandeis University, the sponsors of my laboratory's research, or the companies i am involved with, Abuzz, Xilicon, or Thinmail.
Humbly yours,
Jordan Pollack
Bigname@scientist.com
P.S. you too can be a scientist thanks to mail.com:)
So, you think that we should turn off this Computer. Do you realise that what you're suggesting would be murder ?
I have been plugged in..... I mean alive for about 20 years and can safely say that there are no such things as computers that can log on to Slashdot and post comments!!! I mean this is ridiculous, sure I...., sure there are computers that that can read and perhaps some that are doing basic areobics, but AI, come on! To summarize, I am not a computer planning to take over the world by putting screen cleaning fluid in your drinking water.
I'd also like to add my (sadly, as yet nameless) voice to that of the many others who have praised Jordan Pollack's responses as a demonstration of what Slashdot can be like at its very best.
Unfortunately, in these economically overheated times, when every fashion magazine and gossip rag has a "tech" column, it is all too common for people to lash onto every new gimmick and widget as an instance of the amazing rate of "progress" being achieved in the IT industry, when the truth is more mundane and more depressing: very little of the "new" stuff that comes out of this industry is of real interest, and very little material that is of interest is all that new.
This is not an attack on IT practitioners per se: it is simply a manifestation of the truth that there are no free lunches. Groundbreaking ideas are by definition rare and beyond the capacities of most people to come up with. If ideas as portentous as the General Theory of Relativity or the Lambda Calculus were arrived at every day, they wouldn't seem nearly as revolutionary as we take them to be.
What progress there has been in the IT industry over the 25 years has for the main part been confined to the hardware side of things. One could make oneself out to be a technical visionary in this business by bringing to the marketplace ONLY those research innovations that have occured in the period between 1975 and 1985 - and many, including Steve Jobs and "Uncle" Bill, have have been having brilliant careers of late doing precisely that - so conservative is the mindset of most developers and end-users of IT products. We all love "innovation" in this business, just as long as it isn't too innovative in any way that really matters. Bring us your skinnable web browsers and iMac ripoffs, but those academic pointy-heads better keep their newfangled, abstruse concepts to themselves!
While the web has tremendous promise as an enabler of new forms of collaboration and learning, these alternatives have been, for the most part, subverted by traditional mass-market notions, driven by hype and a flawed perception that capturing the maximum number of "eyeballs" is what web-based publication should really be all about.
Precious few have stopped to consider that sometimes less really is more, and that it can be better to have an audience of 10,000 intensely loyal, actively participatory, highly intelligent and influential individuals than 1,000,000 passive, uncommitted "shoppers" in search of a temporary fix for the boredom that is inevitable in a life driven (if one can ascribe such a term to those who are content simply to drift thoughtlessly from one day to the next) by unexamined conventions.
Judgement and good taste simply are not amenable to democratic sensibilities, and the price to be paid by users of web sites whose owners live in the thrall of the phrase "Eyeballs, Eyeballs, Über Alles" is an endless diet of mediocrity: nothing particularly atrocious ever comes to the fore, but nothing brilliant is ever put forward either. What one gets is the world of iVillage and Maxim, a world in which sex-obsessed Salon passes for highbrow publishing.
For all the annoyance caused by trolls who abuse their anonymity by filling discussions with noise, there is something to be said for Slashdot's review system, at least where Q&A sessions are concerned. Looking at both this Q&A and an earlier one with Bjarne Stroustroup that still remains in my thoughts, I feel my confidence in the intelligence of at least some small portion of my fellow humans rejuvanated: there really does seem to be Intelligent Life out there somewhere!
(1) Red Hat have a ton of IPO money and can afford to operate at a loss. Your favourite band doesn't and can't.
(2) Your favourite band write their own music, they don't have volunteers write it for them for free.
(3) This "give-away-the-software-and-sell-support" thing is a myth. No one is making a profit at this game. Why give away software and sell support when you can lower your operating costs by not giving away software, and selling support ?
I;'ve seen a lot of your posts and this is just the kind of dispossesion mentality that he talks about. You want something for nothing and you want to get it by finding clever ways to circumvent compensating the author, usually by advocating schemes that make it easy to freeload. Can't you find a scheme that doesn't look like a thinly vieled attack on the authors right to compensation ? ( I guess not -- because that's kind of the point )
But what rights will these AI's have? Is it ethical to turn off an AI like this and in effect kill a higher being?
Today most people consider animals as lesser beings that can be used by humanity. Should we consider us humans inferior to a higher intelligence?
I could go on with a long list but you get the general idea. Also for a more fun approach to these questions watch movies like Blade Runner, 2001 A space Odyssey (When HAL kills, who is responsible?) and of course The Matrix (And agent Smiths feelings for the human species.
Well, first I'd like to address your "engineering rule"... NO! This is one of the most evil ideas fostered by modern society that if something is at the extreme its wrong. You drop water to 32 degrees fahrenheit and it freezes. period. An extreme. You are on earth a large planet and you open your hand. Whatever your holding falls. An extreme. The laws of reality are extreme. They are the laws of the universe that engineers must follow. If they say "well, the idea that the mass of the planet is related to its gravitational pull is way too extreme, I mean, NO EXCEPTIONS?!" well then whatever they build falls apart.
And I don't think you'd want to hear him speak about philosophy... I would almost bet money that he is an Objectivist. He would be burned at the stake if he spoke on Slashdot. However, he is right. So if you're interested in truth and you're not afraid of someone telling you that charity and the open source movement are as evil as enslavery and murder... well, go grab "Atlas Shrugged" by Ayn Rand. If you don't want to jump right in, try reading "Philosophy: Who Needs It?" by Ayn Rand. The second just explains to people why they *HAVE* to define a philosophy for themselves. Everyone has one. It determines whether you'll bother to eat something when your hungry. It determines everything that you do. If you don't sit down and think about it, you still have a philosophy... however it is an inconsistent philosophy made up of platitudes and bumper sticker slogans with a few soundbites thrown in...
E.
I extremely enjoyed his discussion of how the open source movement is taking away ownership, I wish I were as eloquent as he when trying to explain the objectivist position... and I wish mor epeople would study the history of collectivist movementsm, especially the beginnings (so they can recognize the ones starting up now) and their endings (so they know to stock up on ammo).
It's just that when you get older, much older, and you see the Nobel success probability function receding as you slide inexorably downward, you realize that without a dissemination strategy you are in imminent peril of defunding.
Prime feature of intellectual evolution.
AI's :Cant live with them, cant shoot them. :How many AI's does it take to change a light buld? 2, one to send an email to the helpdesk and the other to continue the attack on humanity. >'o'
Christ. The worst part of geekdom is the tendency to be flippant and smug when confronting non-technical questions. This is just what Jordan just did re: the ethics of AI. Here is a clue: Resorting to rhetorical questions ("Can a spiritual quest be unethical?" Um, yes...), ad hominem attacks ("Joy's angst must reflect the Sun setting on any instruction set architecture besides x86..." What? This flip response totally fails to refute any of Joy's points), and finger-pointing ("Talk to me about ethics, when.." yadda yadda i don't have to get my house in order if yours isn't) are not legitimate responses when someone asks you to defend the ethics of your research and goals. I agree with his point about the threat of environmental externalities from large-scale industrial and agricultural processes, but YO! Jordan! THIS IS NOT WHAT WE ASKED YOU ABOUT. The fact that driven, brilliant achievers like Pollack seem pathologically incapable of discussing the ethics of what they do makes me alternately concerned and furious. Mostly furious. It would've been nice if he'd answered the questions he was given, or even these: 1) Do the coming ultratechnolgies (like AI) threaten to put so much power into the hands of anyone who wields them that they are unethical to develop as things stand now? (This is Bill Joy's main critique, as I see it) 2) Do you agree or disagree with thinkers like Moravec and Kurzweil who believe that artificial intelligence and robotics will infiltrate and eventually completely replace biological humanity? 3) Do human-level AI's deserve some level of human rights? (As absurd as this question usually seems, I think in this case it is quite valid. After all, Pollack states that he believes thermostats have some dim mental-state equivalent, and his approach to AI relies in large part on harnessing the same mechanisms of self-generating complexity that created our own minds. He's not programming a toaster, here.) If not, why? It's this sort of first-we-achieve-the-goal, THEN-we-worry-about-the-consequences thinking that's lead to kinds of industrial problems he nods at in his interview. I suppose, too, Slashdot itself deserves some of the blame, since these questions were in the pool and just didn't get moderated high enough. Hey. Ethics really DO matter. People die because other people don't have the courage to make hard ethical decisons. Unfortunately, Pollack must have some different defintion of death than I do, since he thinks "we will die quicker from e-mail spam caused by viral marketing customer acquisition schemes". No, I'm not humorless. We just deserve more than that. The whole world does. Thanks for nothing, blowhard. Dan "In Too Much of a Hurry To Log In" Crash
You might also want to add to that list: Paradigms of Artificial Intelligence Programming / Case Studies in Common Lisp. Also by Peter Norvig. A very well written introductory book in both AI and Lisp. (Even if your a hard-core C coder, some experience with a wholly different language like Lisp will be very good for your overall programming skills.)
Object-Oriented Common Lisp, Stephen Slade, Prentice Hall
I started with this book too and while it is an excellent reference book I didn't like it much to learn Lisp from. I'd advise using the book by Norvig I mentioned above or ANSI Common Lisp by Paul Graham (the man who wrote store.yahoo.com in Lisp and then sold it to Yahoo for $48 million) if you're interested in learning Lisp. There's also a nice on-line book by David B. Lamkins: Succesful Lisp
(LISP is the language of symbolic programming, and though I'd rather do my stuff in C*, it does cater to AI programming.)
I'd prefer to do my stuff in Lisp ;-) Just to help a common myth out of the world: Lisp is not just an AI language. It is a general purpose language useful for pretty much anything.
How much does it change your perspective _knowing_ that there are artists in any given field who will eat dirt and ramen noodles for years, decades, if doing so lets them create _their_ artwork? The complexity limitation does not need to be overcome because people have _always_ been prone to go to extremes in their hobbies and arts. A madly romantic, obsessed hobbyist is better than the most highly paid hack. The complexities professionals deal with are far beyond the laymen, but there are always madmen out there so beyond the professionals that it'd give you whiplash just thinking about it- this is the apex of art and craft, where they merge. For once I don't even have to use myself as an example- compare Linux kernel hackers to, say, VBA programmers for fortune 500 companies, or whoever implemented Clippy or Microsoft Bob. Put not your trust in 'professionals' ;) at least, not to the extent that they are your yardstick. This is what concerns me about your argument- you continue to make a case for the big industries of the last century. Even in their heyday, these came down to individual artistic visions. Hitchcock. Lucas. David O. Selznick...
And David O. Selznick is _dead_. The starmaking apparatus he exemplified and personified needs to die now too. It's doing nothing but Pokemon and Ricky Martin, Hootie and maybe a bit of Hanson just for giggles. That's where the safe money is, and it's a crime against art to reduce it to strictly "is this the biggest monetary reward at the end?".
Naturally there will be exceptions, but I am far from convinced there is _any_ justification for assuming the necessity for huge sums of money, the 'blockbuster' movie or 'top 40' song. Who _needs_ to see a blockbuster movie or hear a top 40 song? Who would really be harmed if these things vanished and media became much more localised and focussed in on particular areas of interest? Is there even a need for a full-length movie? If the answer is 'yes, the length is important' then why not a 27 hour movie? The forms media has fallen into are impossibly stereotyped and restrictive. They will crack... and shatter...
I do agree that the majority of groups, bands, or backyard movie productions _will_ fail from lack of discipline. But there are a _lot_ of groups, bands, and there will be a lot of backyard movie productions. As these things become attainable certain artists/groups will develop cult followings- those are the ones to watch. You just need to learn where to look to spot them, as they will not have the promotion muscle of a stultified industry looking for one big hit to do tonnage on. I'm saying that situation is artificial and doesn't meet consumer's needs very well- it's natural to differentiate more than that. There will probably always be a market for blockbusters as well- even if it mutates into a 'pay 100$ (or 1000$?) to don black tie and go to see a movie' in special luxury theaters.
To some extent you're pre-supposing a scarcity of resources. One of the major points in the explosion of technology is that this scarcity isn't continuing. I can tell you that for less total money than some of you put into your PCs, I as a musician and sound engineer (www.mp3.com/ChrisJ for obligatory examples :) ) am able to produce completed work that, technically, kills a lot of stuff released by the mainstream industry throughout history, and on top of that I have access to types of tools that the engineers of the 50s and 60s and 70s would have killed for.
On top of this I have access to global distribution simply due to my using the mp3 format, which is popular and downloadable, so not even _distribution_ is an insuperable problem.
If I have access to all this for so little money, a level of production and distribution that _used_ to require not only million dollar recording studios and mastering houses but also fleets of trucks and distributors, doesn't that say something about how the rules are changing?
And if I can (through a lot of years of study and work and a certain amount of saving and buying and a lot of mad hacking of equipment) do this NOW, what does that suggest about the future prospects of indie film directors and actors and cinematographers as the technology advances to the point where essentially anything imaginable becomes 'filmable', not for billions of dollars but out of the home?
I would humbly argue that 'consumer-grade' camcorders will eventually beat Cinemascope given enough years- and I will stubbornly argue that, given a world full of people to draw on, there is going to be a _lot_ of stuff out there more interesting than backyards and dogs. To suggest otherwise is to suggest that only through media cartels can real Art be created- and... *g* well, maybe that _is_ what some people are suggesting, but my God, such a thought!
The big error, the big oversight here is the unspoken assumption that "because you make something, you can find a market for it". That's, unfortunately, bullshit... certainly both in computer software and in music, the market is distorted by bigger players who control access to distribution and promotion. This is so obvious that people don't see it, it's taken for granted that a cartel totally dictates what you can market in these industries. Yet, it presents a nasty barrier for entry, and blocks the use of capitalism by individuals.
I do music because I like to- I like making it, I can usually pick out something of mine that about anybody would like (it varies a great deal- which is a negative, if you're a cartel: I'm not supposed to be a working artist, I'm supposed to be an easily categorisable disposable artist to be squeezed dry in a single burst of major exposure, then ditched). The thing to remember about my music freeness, determination to make _everything_ downloadable without penalty and withhold nothing, determination to cooperate with anybody who wants to use sampling and take bits of my music for their own use without compensation- is that, just maybe, I have a clearer idea of the real issues here than a lot of people who still believe there's a free market for music (or software).
As I see it- promotion is _everything_. Being known, being publically recognized, is infinitely more valuable than trying to coerce a return on particular artworks or programs- because it is the only thing that can stand against the forces of a cartel controlling the relevant industry. If you are going all business, make a little record label and hammer out good relationships with some stores and pressing houses and a studio and mastering house- guess what? You can't force the cartels to give you distribution. You are _completely_ outmatched there. It is _impossible_ to pursue a 'push' model of product distribution in the face of industry cartels- and if that's where you've put your energies, you're doomed.
If you go all 'free' you're pursuing a 'pull' model. How many really gifted Linux programmers are starving right now? By the same token I see 'free music' as the _only_ reasonable choice (even to the extent that I'll happily talk about all that and tip other people off to it- and I'm told by other people doing this that it actually works!). Instead of 'give, and depend on the gratitude of fans' being a dumb idealistic approach, it is actually the only bulletproof approach- because it is a 'pull' model. You _can't_ force people to listen to and buy your music when you're up against cartels- but no matter how outmatched you are, you can _give_ it away and if people are so happy that they want to help you, nothing can stop them! (speaking of which, did you go to www.mp3.com/ChrisJ and www.mp3.com/RFW and download a bunch of free music yet? ;) )
I'm totally, dead serious. In some ways I'm claiming a pretty serious breakdown of capitalism, that it's not possible for an individual to push their product onto the market in the usual business sense. I _am_ claiming that. But I'm also advocating the solution, which is that in an era of such great connectivity, exposure and ability to get attention are far more valuable than the ability to get your stuff _shoved_ on people- when your horizon is as large as the world, suddenly _everyone_ is 'shoving' their crap onto you, and it all blurs and becomes totally valueless and meaningless. Truth becomes incomprehensible when people will say anything to con you ('innovation'), and in that condition the ability of some small capitalist to get their _demand_ through becomes nil. You can't out-hard-sell a planet full of cartels.
But! As Linux has shown, with this global horizon, there's the capacity for little pockets of generosity and no-strings-attached sincere stuff to get attention, even in the face of the constant deafening roar of the cartels and media. Somebody says, "Hey- didja know you can download all the source for linux and hack with it and you're _allowed_ to?". In the same spirit I say, "Didja know you can download _all_ of my songs on all my (buyable) CDs totally free and not only are you allowed to, I'm kind of passionate that you should never _have_ to cough up money for 'em?" It's partly the simple idealism of "Screw this hucksterism, I want to give stuff", but in some cases (certainly mine) there's also "Things have gone so far out of balance that now, not only do I _want_ to give stuff, but it's the only possible way to earn supporters".
And so, off I go- not only producing good music and stubbornly making it freely available, but also publically offering access to the master tapes for anyone who wants samples, and in fact on Slashdot I've repeatedly offered to let opensource programmer types record music in my studio (you get here, I can't do roadtrips) for FREE when I could legitimately charge $75 an hour, easy. I still offer that- anybody wanting should talk to me. I'm also offering free access to the Geeks In Space people. It's hard to know what more I could offer, but if I think of anything I shall offer that as well :) because (a) I get a kick out of it and like being that way, and (b) again, I see a benefit in it for me, a _social_ benefit not a coercing one. The more I can give the more valuable I become and the more likely it is that I'll keep busy, start doing more work with others, to the point where eventually I can ask $75 an hour sound engineering and it'll seem like an incredible bargain- or have 1% of music listeners buy my ($5.99) CD and have that amount to a cozy nest-egg.
But the important thing is that that doesn't come first. It can _only_ happen if I give so well that it's appreciated, if I work hard enough to be able to really offer something.
So final anti-hype- if anyone could go to www.mp3.com/ChrisJ and www.mp3.com/RFW and give suggestions on what _more_ I can do to give something valuable to people for nothing, I'd be real pleased :) I'm starting work on a techno album which will be pretty unusual, and need to fit in some equipment building in the form of a fancy EQ for my drum sounds. More albums? Better terms for offering studio time? Do people want schematics for the equipment I'm building, or the modifications I make to the gear I have? As might be imagined I'm holding nothing back whatsoever- and like it that way. So, anything else people want?
But water doesn't always freeze at 32 degrees F. It depends on the salinity.
(currently testing something about signatures here)
I appreciate the time spent answer our queries, but there are a couple of points I take issue with
The notion that open source advocates are "left" while IP advocates are "right" is severely misleading. The argument has got virtually nothing to do with left/right wing. The notion that there should be restrictions on copying of information is a form of protectionism. The arguments to justify IP ie "protect the creators" seem more socialist to me than the anti-IP arguments (not that there's anything inherently wrong with that). A "free market" is one in which there are no artificial barriers to entry, and the price of an article becomes the price of production. For instance, some socialist countries have laws that make it illegal for someone to do a job unless they are a member of the relevent union. These laws exist to maintain wage levels for people in those industries. IP laws which exist to protect the investment of the original creators are in some ways equivalent.
> Gene Roddenberry died before coming up with the > backstory for how to have a non-mediocre
> society with unlimited replication for all.
This takes us to the crux of the argument, which is whether there can be sufficient motivation for IP creation without IP protection. For the moment, I think the answer is no, and I believe your IP scheme is probably preferable to the status quo. However, in the long run, I think we will move to a gift economy and that this will work even better.
In richer societies, people don't work in order to survive, they work to improve their status. An advertising executive might be prepared to work 60 hours a week so he can afford a mercedes. Is it because a mercedes is so much more comfortable and faster than a Pinto - of course not, it's because of the status driving a mercedes confers on him. Once material necessities are satisified, you might think people would relax more, but they don't because status is a primary motivator.
Of course status has to be tradeable for a gift economy to be healthy, but somehow status always ends up being tradeable. IP that can be freely replicated reproduces much more effectively than proprietary IP, so we don't need to change IP laws (except patents) in order for a gift economy to develop, it'll happen naturally, and it will work better than a market economy, but the two can co-exist indefinitely.
http://rareformnewmedia.com/
i suck at math, and it takes me weeks to tune any batch file of > 100 lines.
I wish I had a nickel for every time someone said "Information wants to be free".
These are my friends, See how they glisten. See this one shine, how he smiles in the light.
...there is another approach called natural selection that has a really, and I mean really, good reputation.
Good judgement comes from experience, and experience comes from bad judgement.
- W. Wriston, former Citibank CEO
>How many people do you know?
"Knowing" someone is not requisite to assessing their intelligence. Knowing something about them is.
>How do you know if they are moronic if you don't know anything about them?
See, you made the same distinction yourself. Listen to your own words! A really straighforward method for assessing someone's intelligence without having to "know" them is this: watch them. If they act like morons, chances are good that they are.
>Let me suggest that the reason you think a good size of the population are morons is
>because of your ego
I suggest, and I don't need your permission, that your motivation in making this post is your weak ego.
>Every time you come across someone who disagrees with you you are convinced that
>you are right and they are wrong
This is getting boring, have you looked at a mirror recently? This is a typical position. It is not limited to intelligent people, to those that think they are intelligent, nor to those that simply aren't. Those that are intelligent can usually make a decent argument for their position without resorting to feeble ad hominem maneuvers.
>Likewise, every time you meet people who aren't interested in the things you are
>interested in, you decide it must be due to lack of intelligence on their part
You do know this guy really well don't you. I guess you just don't know enough smart people. Many of the bright folks that I know relish new ideas, thoughts, problems, and points of view. For them, a good brain exercise feels good. Food for thought is delicious.
Sorry, but you really must not be that bright. See there? I don't know you at all except by your post and I have already made an assessment of your intelligence, or lack thereof. Whether or not you agree with it is irrelevant. However, if you are so motivated, there are ways for you to change my mind. I leave the discovery of those as an exercise for the reader.
Good judgement comes from experience, and experience comes from bad judgement.
- W. Wriston, former Citibank CEO
>That said, you made some good points in your response.
>Some of those points made me Change My Mind.
That's what communication and intelligence is all about.
Good judgement comes from experience, and experience comes from bad judgement.
- W. Wriston, former Citibank CEO
>Idiots breed faster -- they have nothing better to do.
I appreciate their motivation.
The problem is that we have modified the environment so that it is selecting idiots and their offspring for survival(via their greater fertility relative to others) as opposed to killing them.
Sigh, what are we coming to?
Good judgement comes from experience, and experience comes from bad judgement.
- W. Wriston, former Citibank CEO
I believe that the intent here was that EXTREME's are not evil. When you talk about ABSOLUTES, then that's a whole other game. The original post did seem to confuse the two, but the point is still valid. If hold your hand out flat, place something in it and it doesn't float away, then you can expect to win a bet that says it will drop to the ground if you turn your hand over (unless, of course, it's glued to your hand 8^P ).
Speical circumstances and poorly thought-out statements can always have exceptions too them. That's why laws of physics don't say "everything falls to earth", they say that "objects with mass will attract each other over a distance with a force that is inversely proportional to the distance between them." If I've missed a key element to that law, then I'm sorry, but that is an extreem statement. You can't get away from that in anything that can be observed. Theories might state exceptions to the rule, at the EXTREEMS of a continuum, like black holes throwing equations all out of wack, or when you try to use this law in situations where other forces are at play (magnetism, friction, anhesives, cats...).
The big point here is that extreems aren't always bad, and in engineering (yes I AM and engineer), one must make absolute assertions to even get started. Sometimes these absolutes smack you in the face, "Well, sir, I assumed that if the database contained zero in that field, and the equation tried to divide using that field as a denominator, that either the code would catch it, or the OS would. Sorry your ship's dead in the water." Most of the time, you have to assume that something is a given; 2+3=6, assuming that "+" means "multiply the values on either side of this sign". Most of the time, "+" means "add", but you could change the meaning of that symbol if you wanted to, and you can always make an exception to the rule if you want to nit-pick.
Easy. It's easier to make machines smarter than it is to make humans smarter. Try it some time -- be a guest speaker at a local tutoring company or school.- --------------
----------------------------------------
I thought what was discussed regarding distributed
AI is an interesting idea however my thoughts are
that instead of using our processors, the project
should use our intelligence!
What I mean is that each and every one of us could
participate in "teaching" it
He said AI needs 10 billion lines of code, what
better way to do it than have 10 million people
each writing 1 thousand? (Ok, so when you
work it out its still pretty unlikely but
distributing this project over a larger group of
people will make the goal of having AI that much
more attainable).
I'm not sure exactly how the distributed method
would work as I am just a fan of AI and dont
understand most of the technical issues however im
sure some method of "teaching" could be developed
whether it uses natural language or a set of
commands that users would have to learn. Who
knows, maybe eventually speech recognition will be
good enough for us to spend an hour a night
teaching our little AI bot (yes, kind of like the
way your mother used to read to you at bedtime
except this "kid" would have 10,000 mothers).
Heck, this fits right in with the open source
theme. We can have an opensource AI
"program" where all its knowledge is open
source. I personally don't have the programming
or AI experience to get up right now and say i'll
start work on an open source client/server for
this right now, but i sure as hell would if i did!
I know this is half scifi and am not sure how
feasible this is in actuallity but AI _IS_ scifi
and its pretty damn kool so what do you all think?
-Ezra
I've been very happy with Machine Learning, by Tom Mitchell. It's a well-written introduction to a wide variety of machine learning techniques as well as the issues that bind them together.
I think it's worth reading a good intro text on any topic you're interested in. Once you're familiar with the basics, you'll be able to understand the research papers available online, and that's where the real gold is. Even just reading intro level stuff will put you way ahead of most people though.
Or they are wired in from the start, but the mechanisms or connections atrophy from disuse during the crucial developmental phase. There is evidence for low level straight line feature detectors.
I think face recognition works at birth, but some brain features aren't finished yet at that point. The brain continues to develop after birth, so even saying that an ability is or is not present at birth isn't as meaningful as most people would guess.
Ah, but I think that part of the point was that different people seek to maximize utility within different markets. (or to use an ESR term: Noospheres)
i.e. Linux hackers seek to maximize their reputations within the Linux-credibility market, while Hollywood players seek to maximize their status in the Hype market...
Utility theory is a useful theory for examining behavior in a market. It isn't a complete socio-economic model for reality.
I was disturbed at his belittling of the success of the Free Software movement. I don't think that I agree that people that release their work to the community through the GPL get nothing in return as he claims. He attempts to dismiss what they get as being purely egoboo/recognition, but that sort of recognition translates into employability! It seems as though he doesn't appreciate this.
Employability doing what? Oh wait, I know the answer: writing closed source software.
One of the problems with most open source zealtots is that they don't seem to realize that the closed source software economy is what pays the salaries of the vast majority of open source software developers (like myself).
I would be happy to see a system that was a "happy medium", where source code was available so people could fix bugs, and developers would get paid for the work they do. That system doesn't seem to exist in the real world (though Asynchrony seems to be pretty close). Such a system cannot exist if people expect to be able to get software completely "gratis" though. Somewhere along the line, someone has to pay for the development costs.
This "right" of _possession_ is enforced by the military over the right of life, liberty and happiness. That's what's happening in Chiapas now and in many other places around the world where ordinary, decent, intelligent, hardworking humans are beaten, murdered and raped to keep them producing cheap raw goods for export so that our ordinary decent people will work for less so that the uncivilized and barbarous "middle class" can pontificate about perfect free-markets.
I have to laugh every time I see someone bring up people in third world counries who are "beaten, murdered and raped", as if somehow the abolotion of IP laws will make any difference to those people. Yes, there are serious human rights and poverty problems in the world. Bringing it up in a discussion of open source software is about as relevant as saying "do it for the children" though.
Here's what I'd like: a system where everyone's basic survival needs are met, and beyond that, people are rewarded proportionately to their contribution to society. Does open source help bring us closer to such a world? Nope. The "free market" with IP laws (and enough social programs to give people basic survival needs) is a lot closer than a world with no IP.
If I have a copy of a piece of GPL software, I can personally do whatever I want with it. I can even sell it, if I can find buyers (ask Red Hat!).
That's a very big if. Who would buy your software, if they know that once it's released anyone, including their competiotrs could use and redistribute it. Notice that RedHat doesn't buy the software on their CD's. They get it for free. GPL software is, for all practical purposes, "free beer". People won't pay for the beer, if they can get exactly the same beer for free. (well, maybe they would if they've had too much beer to start with...)
I was surprised, almost 90% of 500-600 people who showed up to an ESR talk responded that they did not get paid to develop commercial closed software products. Most of those people write internal code for companies.
Since when is internal code "non proprietary"? And in any case, the kind of code that people write for internal use is generally not useful as open source because it's highly speciallized. Quite a lot of it entails putting together tools made by others, and adding glue code. In other words, custom hacks.
Who pays for the development of that non-specialized "in-between-the-glue" code? I don't want to be writing custom hacks for a living. I want to write well designed, general software that people at more than one company can actually use.
How many of these people (the 90% you mentioned) were "software developers", and how many were sysadmins? I don't want to sound elitist, but I really don't think someone who works as a sysadmin all day would make a good software developer, in general.
Good software developers learn to always think about the boundary conditions, and to solve the generalized problem. Sysadmins tend to think in terms of "how do I fix this one particular problem on this one machine I'm running into today?"
[If it's true that 90% of open source developers are sysadmins, that might explain some of the truly crufty code I've been looking at in the GIMP...]
Only you >10% of programmers are in danger of losing your jobs (and you guys could salvage it if you'd get sponsorship)
Please clarify what you mean by "you guys could salvage it if you'd get sponsorship".
but you get all the press, you guys take up most of the advertising space, and hence think that you're the most important/most populous portion of the industry.
Sounds like the open source zealtots on Slashdot, actually. I'd guess that less that 1% of Slashdotters are open source zealots, but they make 99% of the noise (once you remove the blatant trolls).
I'd wager that 90% of the software that 90% of people use was written by that "10%".
You are right. Not everyone agrees with me. Ok Allow me to rephrase. According to my definition of "Property", information (ie intellectual works) do not qualify as property. I recognize that others do see them as property and try to respect their worldview, however, I do not and will not share that world view, and will only allow my actions to be dictated by that world view as I deem apropriate.
Allow me to paraphrase:
"According to my definition of 'Property', anything I can get my hands on does not qualify as other people's property. I recognize that others do see them as property and try to respect their worldview, however, I do not and will not share that world view, and will only allow my actions to be dictated by that world view as I deem apropriate."
The ability to produce New Good Music is scarce. Once it is produced, there is no scarcity. I have no problem with hireing musicans to play music, or going to concerts etc etc. Certainly they are talented and should recieve compensation for their talents.
Yes, and how's that going to happen? Noone will pay for it if even people who don't pay benefit equally.
Yes, money is a big motivator. However it is not the only big motivator. Most people are not solely motivated by money.
Money is the most practical way to get food, clothing, and shelter.
You are assuming that there are rights to lose. What you are asking for is not rights to a piece of property but the right to dictate the actions of others. If Alice writes a poem and gives Bob a copy, Then Bob gives a copy to Carol, then how dows the Bob/carol transaction effect Alice? For alice to claim the right to stop the transaction, she is claiming the right to influence transactions between other people, that do not involve her.
The transaction affects Alice because the value of what she has produced has decreased. Ever hear of supply and demand? If there's an infinite supply, the price drops to zero. That would be fine if Alice spent $0 and zero time writing that poem. But if Alice spent the better part of a year writing that poem how can she survive if she can't get paid for it? In an ideal world, the "cost" of producing the poem in the first place would be divided among the people who benefited from it, proportionately to how much they benefited from it. Right now, the free market with IP laws are the best approximation of that we've got.
Consider software development. Thousands of man hours (of highly skilled people) go into producing your typical piece of software. Yes, there is 0 replication cost, but what about that up-front cost? Who pays for all of those developers, and the equipment that they need? I know you're going to bring up open source, but how many developers do you know who actually make money writing open source? Almost all open source developers have a day job writing closed-source software (myself included).
Does this make publishing music less profitable? Yes it does, it means CD pressers would actually have to compete with eachother. It would mean anyone who can press a CD is their competition.
Yes, it also means that content creators get zip.
Does it mean artists wont be able to make money? Hell no. They can still perform.
Explain how people who create digital media like electronica musicians, computer animators, and software developers can make money from "performances". I can already tell that for software developers you're going to say "sell support". Hint: developers don't do support. Support sellers don't need developers, and so they won't hire them.
There will always be a market for live performance. If that doesn't pay well enough, then they will either need to change how performing works (charging more, giving artists a greater cut) or...shudder...get day jobs
Oh yes, get day jobs. because those damn musicians, artists, and software developers don't provide any real value to society. If they want to create, they can do it on their own time. If they want to eat, they can go scrub some floors or something. It's just too bad that the quality of work produced will go down so much, since none of the content creators would be able to afford to do it full time... darn.
BTW, I really liked the way you totally avoided Kaa's point about contracts there.
I believe people should be rewarded for the amount of value they contribute to society. Captalism and IP laws aren't perfect, but they give a reasonable approximation of this. I'm completely open to better approximations, but I've yet to see a proposal for one that seems feasible. Your beliefs almost completely remove the ability for creators of information to be rewarded for the value they contribute to society. The only motive I can see behind your beliefs is greed: you want to be able to get content gratis (for free, as in beer). You don't seem to care if someone spent months or years working on something. I guess you just figure they're a sucker who made something you can use for free.
So what do you do for a living?
No. I'd like to talk to someone who's familiar with the issues.
However people do own things. I can buy something and I then have that something. Nothing changes just perceptions of it.
*smirk* Unless it has "digital content" in it. Then you go to jail for calling it "your" property.
I am not in engineering so I don't know. What I think is the case is that most likely a workable solution that can allow for changing values of various elements would be best served with such an approach.
I'm not an engineer either, I just like the methodology. Another good quote: "better" is the enemy of "good enough".
I like a good philosophy debate. However I would love to know more about AI without being a professor in it.
Me too. 'twould be great to take a class/lecture or have lunch with him. But, given the choice between asking him a question about philosophy or AI.. I'd take philosophy. Afterall.. a few years in the field and it's enough to make anyone believe in God.
Bah. That's impossible - remember the bell-curve? 69% of people are "average". That leaves only 30% on each side of that. Well.. 15% are below, so that leaves 15% who are "above average". Of that, only about 10% could be considered "smart" (again, dividing by 69%). So only about 1 in 10 people are "smart".
I say we need to focus on AI so we can replace the 15% of below-average people who have driver's licenses with robots who can drive better.
Whoah. That paragraph alone deserves an article. Geeks, in their attempts to free themselves from society, gave away their work for free, not realizing it's almost the same thing the other side is clamoring for... nobody owns anything anymore. Isn't it a rule of engineering to mistrust extremes because the answer usually lies somwhere in the middle?
I wish we could have had an interview of you on philosophy instead of AI.. it would have been much more interesting..
less incentive to write/compose/produce, at least as far as financial rewards go. The things that ppl will do for a home in Beverly Hills!
try { do() || do_not(); } catch (JediException err) { yoda(err); }
That was one hell of a Q and A, probably one of the better pieces of slashdot I have read in many months.
AI has always been, to me at least, more about the trip and less about the actual goal. While having these bots of thought will be cool, and yes i think they will come to be, watching how they are being "created" is an amzing decades long story that is unravleing before our eyes. We live in interesting times.
I was also happy to see someone speak thier mind without fear of being whacked by the PC stick of the new millenium facists who sieg hiel under the OSS flag.
Open source is a great thing, what is springing up around it (ie the sceen of closed minded slogan chanters) is far from helping it become all it can. (go read any Jon Katz piece to see where slogan chanting and short attention spans wind up)
Money is a valuation of time, it is also a valuation of labor. At its simplest it means I can translate an hour of working on a dB system into some food, a pda, some books, a gift for the wife and kid, and then some cash to donate to a charity who in turn can feed and clothe and house others.
I have a worth that my employer pays for. I find vlaue in food, pda's, books, giving gifts to my wife and kid, and in donating to others the means to be fed and clothed etc.
That little sqaures of cash with dead presidents is the token of exchange is not the point. There have been works on passing valuation in the form of bits,and to some respects this is happening.
Those who think a world with out captial or value exchange would be a good thing need to come up with some hard and fast examples of Actual Working Models. I can always read the fantasys of Mann and the like if I want pure paper postulation. But that doesnt feed my family.
Thought experiments are a great thing, they can shape the ways in which we better our world. Imposing the pipe dreams of the few often leads to much harm. You can be the next POl POt without even realizing it.
Poor little clams! Snap! Snap! Snap! Poor little clams! Snap! Snap! Snap! Poor little clams! Snap! Snap! Snap!
Sure, technology evolves rapidly. What is state of the art today, will be consumer grade tomorrow. However, the fact that it is technically possible to create a hollywood image quality film, does not mean society will ever view it like a hollywood film. Fifty years ago, on hearing of today's internet/computer/printing technology, one might have made similar predictions with novels, that it would obviate the need for these large publishing houses. Yet, when you look around today, this is clearly not the case. The popular authors for the eighties and nineties (e.g., Ken Follet, Grisham, Clancy, etc.) are still largely the same. Clearly, the technology for internet distribution exists. Yes, the facilities are in place (except if you regard decent "e-books" as being necessary). Yes, it is fully possible. Yet, it just isn't true. The publishers are alive and well, and, I feel, will stay this way.
You are ignoring issues such as marketing, distribution, signal to noise rations (e.g., Any idiot can get a word processor and write, but few can match Grisham's published writing. How do we pick the diamonds from the rough, so to speak? how do we develop talent? who pays for them to work at it?), professional editing, full time employement to produce quality work, etc.
Even going beyond toss-away novels, I'm hard pressed to find any areas which the average Joe has really replaced the professional. The internet has been around long enough now, we (you) would expect it. Sure, it brings new stuff to the table. Such as perl, linux, etc, but they are few in number, and none of them really compete head to head in the same way. They're simply alternatives. Perhaps worthwhile and notable, but not earth shattering. The "democratic" internet has yet to demonstrate otherwise.
Ummm, the fundamental difference between roads and large scale public subsidizing of the arts, is that with roads there is a very clear and tangible public benefit. Few sane people would argue that they would be better off without roads. They are willing to pay relatively small taxes on something that fundamentally benefits them. The music industry today makes billions of dollars off of consumers. Every consumer who wants to listen spends money. With your model, in order to produce approximately the same quantity and quality of music, the government must spend atleast the same amount on the aggregate(e.g., 100 billion a year, or whatever it is). Are you ok with this level of taxation? It might very well quadruple your taxes if government is to subsidize your entertainment. And don't think for one minute that the government can replace free markets well. In other words, if a band sucks, few people will listen to them, resources will not be wasted on them. With the government model, we're going to require some buearacrat or some process to "decide" how to allocate resources. How do we differentiate truely "good" music from "bad" music with government? What socio-economic groups are to command the lions share of the music types? Is this to be done by population?New talent will get discovered how? Where is the incentive to take a risk? ....So many holes...
Never mind the efficiency issues. Or the secondary and tertiary economic effects of high taxes....
For those wanting to find out more about the economics of intellectual property, I recommend the this analysis. The piece is about standards, but the first two sections give a very helpful discussion about the economic rationale for and meaning of IP law.
This man is probably the single most captivating and interesting person slashdot has ever interviewed... I'm not sure what I was expecting, but that's not what I got.
:)
... But find a great graduate program in computer science, and you will likely find fun and exciting work for no salary and no equity!
... participating might provide less utility.
... somebody else had grabbed the URL!
These excerpts in particular made my jaw drop:
I think we have begun to understand the nature of mediocrity as an attractor in educational systems and how to change the utility functions to avoid collusion, and apply this to human learning (Elizabeth Sklar).
I am on a spiritual quest to understand [God as] the principles of the universe which allow self-organization of life as a chemical process far from equilibrium which dissipates energy and creates structure that exploits emergent properties of physics.
[re: Turing Test] So I propose using the Louis Armstrong Test, his answer to the question "What is jazz?"
Reading is fundamental.
Talk to me about the ethics, when your very own open source movement leads to the inevitability of an Intel instruction set monopoly by providing a useful alternative to Microsoft
Jobs in AI are just like software jobs everywhere:
I think a SETI for coherent intentional behavior emerging out of the infrastructure would be a fun project...
If RMS was a radical advocate of anonymity who wrote the GPL so you couldn't put your name on the source code
When I recently tried to corner the market in pig latin domain names
Once secure distribution of programs on a rental basis is established, all content publishing will move inexorably into that mode to maximize profits.
... it is the soft underbelly of "Starship Economics" that Gene Roddenberry died before coming up with the backstory for...
-- What you do today will cost you a day of your life.
Well...
The use (and acquiring) of information depends on expenditure of time, so it is also a limited good. Just to be nitpicky.
What we have here is a matter of degrees. Unfortunately, people don't reason well intuitively about large or small magnitudes. Micro-credits would make sense here. Total control doesn't. But that's not the way to expect laws to be written. So the best choice is something which *totally* protects the individual rights, even though a truly fair balance would tilt things a little bit the other way.
The paperclip is made of steel using the production line, paper money, and millions of other snips of prior art. But a patent on paperclips give one "total" control over their creation. (For a limited period of time, true.) This doesn't acurately reflect the degree of contribution that invention of the paperclip involved. OTOH, the electric lightbulb required considerably more development. And requires even more support. And many inventors aren't suited to be businessmen. And....
For the creation of a snippet of software there seems to be no legal method for the creator to be justly rewarded (though for larger creations nagware is pretty close!). GPL seems, to me, to be the best of the available alternatives.
I think we've pushed this "anyone can grow up to be president" thing too far.
I feel like a better person for reading this column...
I want to have that dudes children
"What do you do with the mad that you feel when you feel so mad you could bite?" - Mister Rogers
The essence of intelligence is performing actions appropriate to one's objectives given limited computational resources and uncertain, incomplete information.
-Devika Subramanian, AI researcher
Now, now... it's not nice to talk about your government like that... :)
--
The best way I've heard it described is that once we have acheived something we considered a step towards AI, it no longer appears to be AI, and we change our definition to look at the next unsolved problem. There is that long range ideal of a sentient machine, but even the definition of sentience is subjective...
--
1900 - my gun is bigger than your gun
2000 - my nuke is nastier than your nuke
2100 - my AI is smarter than your AI
:-)
Actually, I'm not totally convinced that having more "free" software is in the same category as having more "free" beer/money. Firstly, by default all IP whether patent or copyright passes into public domain after 20 or 90 years. So in essense all ideas/thoughts are already "free" (if you take the view that copyleft is merely a time-shift into the future but looking back), just that some people want to create a little (OK, a big) profit opportunity in the short-term. Also ideas breed and mutate on themselves and thus having a pool of software that people can tinker with creates new opportunities (see GIMP and online photo lab). There is no extra cost in having 1 or a million extra people play around with ideas or information. On the other hand, anything to do with atoms is a wasting asset (depreciating, opportunity costs, dead-weight loss, etc). Having an extra 1 million people means that it ties up a lot of resources. In that sense, money is a claim on future resources and the changes in value of various classes of goods, an indication of what the larger population wants, ie signals to suppliers to invest more in certain lines of production. Creating false and/or distorting signals through funny money tinkering of the ol printing press (inflation) is thus a really bad idea (ps someone tell Greenspan to rein in the loose monetary policies). In this sense, OpenSource software is a luxury good, it can only be produced by highly skilled/education people (ie spare cash + food to put themselves through uni or self-taught) with leisure time on their hands. While it may appear rather chaotic to corporations with top-down control, perhaps it is just as efficient (or equally inefficient) as trying to charge everyone for the priviledge of access (ie supply driven rather than demand driven).
Human nature is funny sometimes, especially when people confuse price with value. Cows won't eat mouldy hay but if you put a fence around it, they are tricked into thinking it is forbidden and therefore must be "good". Thus we have a stupidity tax (not knowing where to get hold of free sources) plus an ego-tax (trusting the overhyped brand) which in the long-term, will only benefit the smart hackers. It's regretful that corporations have seized upon these little idiosyncracies of human nature (not to mention doing their damn best to reinforce it) and are making excess profits but then that is the nature of market evolution. Eventually the next generation wises up and the rules change yet again. Having AIs may tilt the balance somewhat in favor of the consumer (imagine having a guardian angel that informed you if something was a good choice or not, e.g. calories in a cake) but then the companies would come up with the devil's advocate to just be daring and try that scrumptious chocolate cake. A never ending arms race, even with AI.
The more things change, the more they stay the same.
LL
This takes us to the crux of the argument, which is whether there can be sufficient motivation for IP creation without IP protection.
Let us not forget those of us who create for the sake of creating. On my webpage for example, I have posted many of my writings and poetry. I didn't write them to make money, I wrote them because I had something to say. Anyone can downlaod and print them out. In fact, bascially the only restrictions I have are that I be given credit for them and that my permission is necessary for commerical use. The motivation for all IP is not money.
Let's concentrates on making the PEOPLE smarter, then worry about the machines.
Well, one obvious approach is called eugenics. I has a really, and I mean really, bad reputation.
Kaa
Kaa
Kaa's Law: In any sufficiently large group of people most are idiots.
...there is another approach called natural selection that has a really, and I mean really, good reputation.
Idiots breed faster -- they have nothing better to do.
Kaa
Kaa
Kaa's Law: In any sufficiently large group of people most are idiots.
Sigh, what are we coming to?
See Kaa's law.
Kaa
Kaa
Kaa's Law: In any sufficiently large group of people most are idiots.
Monty Python
Strap-on brains for stupid people!
If movies could be distributed freely, you would make money by:
Advertising on the server hosting your movies. Other servers could of course host the movies - you would make sure then that your server was the most appealing (better local replication/more bandwith/etc.)
Theatre profits. Even if a movie were relelased the same day it was released in the theatre, many people would still go see the movie. In fact I feel that MORE people might want to go see a movie in the theatre if they could watch a bit of it first... there have been many movies I've watched on video that I wish I would have seen in the theatre but just didn't think it would be good enough at the time. This is also the whole reason behind region encoding (they do not want people to see a movie before it hits the theatres in other countries) and in my mind is foolish and loosing the movie companies a lot of money.
Merchandise. Admittedly, this only applies to action movies and kids movies - right? Wrong! If a movie is good enough people will buy things related to it - imagine if you will "The Ususal Suspects" action figures, or pewter figurines. People WOULD buy these things.
Video distribution. Even if you distributed movies for free, you could still make a bundle if you sold the movies at a fair price on a physical medium like DVD. Lots of people don't want to bother burning a copy out or having to store it - plus it's nice to get a case for something with good artwork.
---
Already the video distribution and merchandising of movies give a far greater return than the theatre runs. So this only makes sense, but movie studios are too stupid/greedy to really make the kind of money they could be by giving the movie away as a "sample" of sorts to promote other goods!
"There is more worth loving than we have strength to love." - Brian Jay Stanley
utility is what motivates you to make decisions in your own self interest.
What appears to be missing in your example is concept of a unified ego. *IF* a person has a unified ego (i.e. no neurotic self-limiting), then you can perhaps talk meaningfully about utility. When a person does not have a unified ego, then there are *two* functional foci: what he consciously wants and what he unconsciously wants. The two utilities conflict with each other, and nothing is accomplished. However, the fact that no progress is being made is a sign that meta-awareness is necessary. The new, meta-aware -- consicous -- utility is that the problem of frustration must be solved before a unified focus *can* exist. So utility *does* enter the picture -- it's just that the concious focus temporily changes from "being frustrated" to "setting a goal of solving the problem of self-frustration".
Of course, since being unconsiously frustrated largely stems from fear, *that* is the problem that must be solved, and it can only be solved *consciously*.
This reminds me of the old joke: How many psychiatrists does it take to change a light bulb?
A: Only one, but the light bulb has to *want* to change. (*rimshot*)
DNA is a Turing machine. You, however, being dynamic and emergent, are not.
....and a lot of creative people spew forth what is often called "art" or "music" or "literature" not out of any conscious motivation, but simply because they have to. The fact that others become aware of their creations can be as an interesting after-effect.
So...."marketable" human creative output will continue regardless of other changes, as long as folks are folks.
Well, one obvious approach is called eugenics. I has a really, and I mean really, bad reputation.
/hist102/photos/html/1150.html
http://us.history.wisc.edu
There's always so much griping on these threads I thought I'd just say that I think this piece is Slashdot at its best. A well thought out, well linked article by someone who actually knows something about what he's talking about, and someone I wouldn't normally be able to hear from.
Worth another month of Trolls and MicroSux posts for sure.
Hotnutz.com - Funny
This sounds like a demand to me, not freedom.
Programmers are not being talked 'out of their writing'. They are being talked out of oppressing others. There is nothing weird about selling free software and open source programmers are just a capable of selling their writings as a closed source programmer. The programmer whose only motivation is money should simple not program unless someone is willing to pay them. If someone comes along and says, "Hey, I'll pay you to write some software for me." Should the programmer care what licence the software ends up under? No. Actually, the programmer should prefer an open source licence in order to maintain the use of the software should the programmer give up the copyright. If the programmer does not give up the copyright, the person paying for the software should demand open source. Lots of programmers today make lots of money writing open source software. They have not given up anything (except the right to oppress others).
just raise the bar for average intelligence, which is a lot higher than most /.'ers seem to think all ready. Why build a new machine when you already have one to work with that has gone through millions of years of evolution. Even the interviewee, after all his years of research, has concluded that the only way to build a good AI is to build one and let it evolve.
--
+&x
Jordan seems to think that the system needs to be based on ownership. I would disagree. I don't need to own a song in order to enjoy it (e.g. radio). It's the person releasing the product that requires ownership and scarcity. So how do you have a system that allows ownership for the seller but not for the buyer? You lower the cost of the product (because it's not resellable, its value is less), and you increase the number of people "buying" the only copy in existance.
How many times have you said "I'd pay to verb that"? You wouldn't pay much, but you would pay. With sufficient demand for a product, that prodcut can be released to the financial benefit of the producer and the general benefit of society.
This system would allow me to write software for a single company, which really needs it, and sell it for some large amount of money, and at the same time, write a generally useful piece of software, and sell it for pennies to thousands of people.
Just as a curiosity, what was the "utility", Jordan, gained from writing for /.?
(flamebait intended oldie)
What do you say to the people that feel it is unethical to try to create "intelligence"?
>I take this as a shorter version of the longer religious question the editor thankfully didn't select.
How about taking it in another way. If AI can do some good for the world, it can do some bad as well...:-( AI is potentially very powerfull tool (IMHO more powerfull than atomic weapons).
>Jobs in AI are just like software jobs everywhere
Who says AI is to be made in software... I sure don't think so. I think AI is about communication, and that (amount&kind) just isn't available on any current computersystem.
>but nobody can write a coherent program bigger than 10M lines
Which proves my/another point: who says it needs to be one single programme...
nosig today
Mucho respect to Jordan Pollack. He has obviously spent a lot of time and thought responding to these questions - and the number of links!
The way he has answered these q's shows he has respect for the geek community, and in particular has done Slashdot a service by providing some quality content (not that we don't get it normally), but makes a better interview than some special guests in the past (recalling one nutter who refused to reply properly).
"At a risk of being called sexist, ageist and French, if you put multimedia, a leather skirt and lipstick on a grandmother and take her to a nightclub, she's still not going to get lucky."
- Jean-Louis Gasse
insignificant sig
That's not actually tautologous. Implicit within that statement is the assumption that there is a something which makes people do things. Strage as it may sound, there have been psychological schools (or, at least, school) of thought which challeneged the idea that we actually make conscious decisions of any kind. Skinnerians (IIRC) argued that it's all just post hoc rationalization.
I think they're wrong (and kind of loopy) nevertheless, it does point out that there is an assertion of value in the statement "There is something which motivates people, let's call it utility."
I think you're getting hung up on the idea that calling it utility means something. That calling this motivational attribute "utility" instead of "quality" or "value" or "vorgelspatchny" or any other word imposes some limitation on what it can be. That because it's "utility" beauty, or pleasure, or so forth don't count.
But that's wrong. "Utility" here does not exclude those things (or any other), it embraces them all. Don't get hung up on the word.
What the assertion really boils down to is "people do stuff for reasons" -- which is in contradiction with Skinnerians and some other folks, and thus is non-trivial -- "and those reasons are subjective." Nobody is suggesting those reasons are trivial, simple, easy to discern, non-multiple, never-conflicted or rational. "The heart has its reasons..." &c.
It boils down to "Ya do stuff cuz ya want to." Which is not a tautology at all: it's an important idea about how the mind works.
----------------------------------------------
-*- Any technology indistinguishable from magic is insufficiently advanced -*-
.. and then consider that 50% of people aren't as smart as that.
If this was idea was followed then we have the problem that we would be in a massively inflationary state permanently, We can see that a significantly large part of the national economy is tied up in software licences. (if not how are Microsoft making all that money.)As this years $100 computer program turned into next years $50 program, Moores law would be hell on the economy. for every percentage of the economy that is made up of software licences you would end up with 1/2% devaluation of your economy. So if the Information industry made up 10% of your economy you'ld be talking about 5% devaluation of your economy and a masive spiraling of inflation.
We would also have the problem that to preserve the currency as strong and stable the only thing that would end up happening is small scale software companies would be shut down as their efforts could damage the national economy, by rendering a piece of software obsolete. Open source developers would then be viewed as economic Terrorists.
Maybe we're taking the wrong approach to AI atm. Maybe we shouldn't try to create something with human intelligence, or even monkey-like intelligence. We should aim small to begin with, then build on what we discover.
AI and A-Life are a big hobby of mine... I'm developing a software system that should be capable of evolution (in the Darwin sense (survival of the fittest), not the "oh, I don't work, must fix myself" sense). However, finding an environment where a piece of software has netural enemies will cast doubts onto the nature of the system (can you guess what it is?), but with any luck, I may get results...
- Damnit, I'm dead Jim
"Let's concentrate on making people smarter and then worry about the machines."
Well now, a truly fine idea. But how do you propose to do this exactly? Throwing more dollars at education is very much not working. Perhaps you could shift cultural values (as this seems to have more real effect) but cultural control knobs are notoriously difficult to find and manipulate. And what about those folks who believe they are as happy as can be in their abysmal ignorance? You can't exactly force them to learn.
At the other end of the scale - I don't know about you but I am very aware of apparent limits in my ability to learn and to reason clearly and well. I don't think these are just my own personal short-comings as I actually measure pretty highly relative to humanity at large. I think this model of humanity has some real mental processing limits (without even getting into psychological limits for now) that are more than a little problematic as the world gets more and more complex. Frankly I see very little reason to believe that any human or set of humans is adequately equipped to deal with some of the issues confronting us well. And this is after watching various proposals and essays and systems being built for quite some time.
So, imho, we had better hope we can build true AIs and/or seriously augment our own intellectual abilities. It may well be our only chance.
Yes, this month's Wired gives a good summary of the history of MT. It gives the firm impression that MT will remain several years off for quite some time. And yet, we will continue to see useful, special purpose applications of what is learned in trying to achieve it. The article emphasizes that MT has had its greatest success within constrained domains. It even mentions that Babelfish, for all its warts, is particularly good at translating recipes.
The net will not be what we demand, but what we make it. Build it well.
What scares me about this is that programmers learn early on the real meaning of GIGO: Garbage In, Garbage Out. Regrettably, in the real world, the garbage that goes in (sometimes the data, sometimes the assumptions on which the code is based) often remains unexamined. And the output is treated as gospel. In particular, with data mining, there is a crying need to educate the users of the output about the potential for errors in the input, bugs in the code and just incorrect assumptions (e.g., I bought one book in a series, I must want the rest). Data mining is a fine way to narrow down our guesses, increasing the odds that we are right. But it is not and never will be an exact science.
The net will not be what we demand, but what we make it. Build it well.
well we've probably just about given up on a whole bunch of ourselves. Being intelligent and thinking and caring about what goes on around us just doesn't rate that highly any more. I think people are more interested in tangibles these days, what you can see is what you have and so enjoy it... People are pretty apathetic about "obtaining" abstract intelligence when they can just learn how to make $40K a year a live happily ever after or sth...
Maybe though, the development of A.I. could really inspire a lot of people to get involved, much like a ton of people became interested in space, space travel & physics in part due to break throughs in space travel, maybe AI just needs that one big breakthrough to bring back the romanticism in it all again... The blissful optimism that anything is possible, today and tomorrow...
On a slightly different topic, the critters we've got running around today would have been unthinkable 5 or 10 years ago. The MIT AI lab face recognizer, the cars that drive themselves (Albiet in very controlled environments) the assorted space probes that can adapt to various conditions... we've come quite a long way without realizing it.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
The various techniques used in AI (feedforward NNs, SOMs, genetic algorithms, etc.) are all useful for a variety of tasks, although they have different strengths. That's the whole point - I can simulate human sound processing using a spectral analyser and library of patterns, but an SOM can adapt much better and do different tasks.
As for nervous systems, they only really have one tool - the neuron. It's the connections between them that make the difference, and while we know that there are vastly different connection patterns in different areas of the brain there is no evidence for particular "tools" being used in particular areas. That's why AI researchers have been trying to use NNs and similar techniques for every task under the sun - the human brain does it that way.
I do think that AI researchers are stuck on some things - like emulating the brain. For example, some models of visual processing take pride in the fact that they can replicate optical illusions in a similar way to the human eye - is that really relevant? There also seems to be an obsession with self-aware AIs, so it's good to hear from an AI researcher interested in any intelligent processes.
I think that's enough rambling from me for one topic.
Something else to complain about: He blew off the one question that i really wanted to see answered: How to get more involved. Reading isn't being involved (although helps)... is he being exclusionary? Screw that...
Mike Roberto (roberto@soul.apk.net) - AOL IM: MicroBerto
Berto
That is just not true. And if you had bothered to read the artical he points out that IA is curently in use. What do you thing handwriting recognition is based on? Just remember next time that you send in you taxes an AI read the address on the envelope and then another read all of the data on you tax form and selected you for an audit!
Has anybody applied the Turing Test to this guy?
Assuming the increase of intelligence is always accompanied by the decrease of entropy Why the hell would I assume that? Any way ever synce "The Fractal Geomotry of Nature" we have had a much better understanding of the relationship betwen compelxity and entropy. And infact the complexity of the univerce increces not decreaces with entropy.
Assuming the increase of intelligence is always accompanied by the decrease of entropy, the amount of Intelligence in the Universe must not be increasing due to the second law of Thermodynamics.
Just a 1 cent thought. Someone please tell me I'm wrong, and why.
> which is a restriction that TheCarp himself
> places upon people who choose to replicate
> his intellectual property at no cost to him.
It is?
hmmm I probably did write that at some point.
Actually...I see nothing wrong with claiming
credit for ones works. To say that one wrote
something which one did not, is fraud.
In the end though, I truely don't care to much
about it (I am assuming that you went to my web
page and saw that statment by me that I want
credit...heh thats several years old. I am too
lazy to update the page. Certainly my personal
beliefs have evolved greatly from the time that
I wrote that)
"I opened my eyes, and everything went dark again"
1 word: education (well, maybe you can add in neural enhancement through genetic engineering/ nanotech as well). You just wait and see..the entire point behid the "information revolution" is to destroy the monopoly that the higher social strata have possessed on knowledge (along with all of the power that necessarily comes with knowledge) for the totality of human history. It is this monopoly--more than any other single factor--which has guaranteed the subjugation of the mass of humanity and all of the attendant social ills with which we have become so intimately familiar .
In 50 yearz, one can at least fervently hope that the 'knowledge scarcity' you speak of will no longer exist. People everywhere may well be more knowledgeable/intelligent/talented than we can imagine.
If, a few decades from know, the average person has the intellignece of a shakespeare and literary and technical resources which he could not have dreamt of (in heaven, earth, or hell), no IP protection will be necessary. A gift economy could rule effectively. I think we would feel better..things would be done for curiousity, love, scientific fascination, etc--not to temporarily allay our insecurity concering the necesities of life or to increase our social status in response to some silly genes for hierarchy dominace/sexual competition left over from our days as chimplings. People would create for the love of the act of creation itself; the appreciation of beauty and meaning brought into existence would be motive enough. These notives are just as universal as, say, greed, and they have never been given a chance. Who can say whihc is more deeply rooted in humanity? (of course, soon we will be able to change even human nature itself according to our whim...)
Sound silly? I repeat: just wait until everyone is smart and not concerned with dying of tuberculosis, war, or starvation. All members of the human race will be capable of altruism and creativity little thought of in these dark times...
There are, of course, many things which could prevent this scenario from becoming a reality, but man, if we do make it.......
The truth is out there - we'll let it back in after it sobers up a bit. -The Cube
To fly, mankind had to realize that there were no need to move the arms like birds. To create conscious machines, we have to stop replicatating the brain or nature.
If everyone multiplies their bank account by 100, the value of those accounts is instantly divided by 100.
I meant: ...the value of each dollar on those accounts is instantly divided by 100.
So the value of the account is constant.
All opinions are my own - until criticized
Agreed, less music might be recorded since there would be no reason to create "IP" just to speculate for financial gain. But as long as there is a demand for music, someone will create it. This might come as a surprise, but music is actually older than recording technology!
There is also a demand for "stars" so there will be stars, just not prefab RIAA stars, but more unpredictable. And *they* will buy houses in beverly hills so I'm not worried about that area either.
All opinions are my own - until criticized
Still a hundred copies would *cost* me a hundred times more than one (except for discounts of course)
Cloning the BS album would be more like selling a dollar bill to myself a hundred times.
Anyway, about your coin collector:
Someone who would spend $25 dollars to buy a quarter from the 1800's. This value increase is based on scarcity, a property normally attributed to IP.
Nineteenth century quarters are scarce, will *remain* scarce and hence has a collectors value. Britney Spears CD's can be *made* scarce. The music upon them is also scarce, if official CD's are the only medium. They get an increased *relative* value, not by creating more/better "property" but by keeping others out of supply.
Scarcity is a property artificially attributed to IP to give it a financial value. IP also has real value. Freevare is not scarce, do you mean that free software is worthless? Do you claim that air is worthless since it is equally distributed?
Interactions are based on the differences between, not the absolute values of, the temperatures.
True, but lukewarm coffee taste like shit, even if I brew it in the freezer.
All opinions are my own - until criticized
Ah, but there is a *major* difference between a hundred dollar bill (in or out of the bank) and hundred dollars worth of software: :-)
Money has no use except as currency. (You could usa a coin as a primitive screwdriver, but that's about it...)
Software, music and other forms of "IP currency" are useful in them selves (even Britney Spears and PowerPoint, just to save you from the obvious reply
If everyone multiplies their bank account by 100, the value of those accounts is instantly divided by 100. If everyone multiplies their music collection by 100, everyone gets a hundred times more music.
All opinions are my own - until criticized
Real AI is here today! Experience artificial intellegence in your favorite editor: just type ':set ai', vi users, and experience the joy of intellegence!
"The romance of Silicon Valley was about money - excuse me, about changing the world, one million dollars at a time."
Visit
Humans are destroying the world that we live on through pollution/war etc.
Humans create artificial intelligence.
Humans die out due to above problems without having populated outer space.
After many eons, AI's get bored and decide to correct atmospheric problems and resurect humans for entertainment.
Humans live once again!
Was this outcome ethical?
Moot!
no sig.
Well, considering that a goodly size of the population are a bunch of morons, I would say that Artificial Intelligence is better than none.
If not now, when?
You should have read the article instead of posting as fast as you could. He states that AI is already here in different forms.
;-)
The little Microsoft Office2000 Cat that answers all my questions is a form of AI (1st category according to article).
Microsoft might eventually have a NT-assistant...
You can select from a paper-clip, office shredder, or cartoon 2-year-tech-degree System-Administrator to answer all of your questions. With a little more work, this NT-Assistant might replace the real thing! For a little realism, the NT-assistant could get frustrated when you showed it a CD that you stuck in the microwave and said that the CD-burner wasn't working.
-
-=RR=-
"Yea, cus when you take that bus; you get there" -Tone Def
Insanity is the last line of defence for the master diplomat. But you have to lay the groundwork early.
I hope that I live to see the day that the internet gains self awareness, because it might look at all the spam mail about US only offers available forUS dollars only via a US only 0800 number and realise it's a waste of time routing them to a .uk domain. Much simpler than tryin to educate spammers imho.
- Andy R.
A pizza of radius z and thickness a has a volume of pi z z a
The statement "utility is what motivates you to make decisions in your own self interest" is not the economic definition of utility. According to _Intermediate_Microeconomics:_A_Modern_Approach_ by H. R. Varian the notion of utility as a measure of happiness (which makes it a motivator) comes from Victorian ideas and is not the current economic view of utility.
Instead it states: "utility is seen only as a way to describe [consumer] preferences". So basically utility is a measure of what people actually choose to do. So Jordan's statement is correct except that utility is actually a measure of what happens. It is not an intrinsic absolute measure of happiness, but rather a relative one used to compare a persons possible choices and predict what that person will choose.
Disclaimer: I'm a EE not an econ major. My econ experience only includes a few courses.
(note: I'm not necessarily advising the book I mentioned. It just happened to be the first one I picked up)
Ok... enough of this stuff. The idea of taking The Golden Mean when faced with a range of choices dates (in 'published' form) from Aristotle, who expounded at great length on this concept and its validation as a method.
(Yes, this counts as "a long time ago.")
The Golden Mean was developed from the Socratic dialectic , a concept which is probably more useful.
[|]
Actually, APL, LISP, and Java can be hard-compiled to machine code. LISP usually is. Java sometimes is; SGI has a hard-code Java compiler. (This is separate from all the JIT compiler stuff.) APL is an interesting case, as there are a whole class of optimizations possible for APL because of the regularity of the language and the direct representation of operations on collections. There's a forgotten technology of APL optimization, although remnants of it live on as massively parallel techniques for supercomputers. Incidentally, Hard-code APL compilers have been written. (There's even OpenAPL for Linux!)
The "language as entire environment" idea recurs frequently. BASIC is the classic example. Turbo Pascal and Smalltalk were also closed world. But with the widespread availability of good integrated development environments, there's less need for such schemes than there used to be. Today, you can write in C++ and have all the ease-of-use facilities, including a source-level debugger, you had in the closed environments of old.
Of course, if you're stuck using GCC on Linux, you may feel you're missing something.
Wasn't the Pinto (one of) the car(s) that exploded in simple rear-end collisions?
In that case, no, your advertising executive is not working 60 hours for speed, comfort, or even status. His rationale is much more simple: "But I don't wanna die in a massive fireball!!"
Next time, choose a better car for your illustrations and examples ;)
I am simply saying that permanent use and resale licenses to changeable information (software, art, literature, music, movies) which can be traded securely, without loss or duplication, in a public market, is a form of currency.
I would respectfully disagree with this analogy. Currency has to be limited because it is a reflection of the value of tangible goods. Increase the number of currency units by duplication and the ratio of currency units to real goods drops resulting in the devaluation of every unit.
However with information such as software, books or movies, their value is not based on any numerical relationship between them and tangible goods, but on usefulness or sentimental (entertainment) value, neither of which is diminished by producing an identical copy. Duplicating currency units affects every other user of that currency. Duplicating software and entertainment does not affect the other users in any way.
--
--
Things are only impossible until they are not.
I've not seen one that has. They model charactures that are exteremely unlike a neuron. If they were serious about modelling a neuron, they would model the sodium/potassium pump. There are certain neurons (Pyramidal, I believe) that actually fire spontaniously without a synaptic jolt. They simply pump so many ions untill a critical voltage released and the neuron discharges. I've never seen any models that take these behaviors into account.
--------------------------------------------------
It's more than likely that many if not all of the different aspects of intelligence are needed to form into an AI.
One case in point are Kohonen neural networks, which are a close model to the behavior of how sound gets processed by the brain. The Situated Action camp has produced models that very well model the motor control and locomotive behavior of insects and fish. And then there is the field of emergent intelligence (related to situated cognition) that seems promising in its attempts to model large scale collective cognition such as ant colonies or beaurocrasies.
Isn't it possible that nervous systems must utilize many different tools that act in concert? I think people should split resarch into the following separate categories:
coyo the coyote
----------------------
new url: http://www.twu.net/~coyo
--------------------------------------------------
What we REALLY need right now is some REAL intelligence. I mean, sure, it's nice to have "intelligent" machines, but what good does that do when a good size of the population are utter and complete morons? Let's concentrates on making the PEOPLE smarter, then worry about the machines.
A genius writes code an idiot can understand, while an idiot writes code the compiler can't understand.
Maybe my point about monolingual programming environments wasn't clear (since I was trying to avoid being moderated as flamebait). The example with Perl illustrates this best. I don't think that Perl should exist (really). Rather, I think that Perl's existence is due to deficiencies in the C/Unix programming environment. And using C++ makes no difference. You still end up with a trilingual programming environment (C++/Perl/Unix), rather than a monolingual one. This is adding complexity to resolve the problems: I believe that makes things worse.
Monolingual programming environments require less programmer training and ultimately lead to reduced software complexity. This is an important point--whether you agree with my views on their overall desirability or not--and it is a point that most people don't consider nowadays.
APL and LISP are Very High Level "applicative/functional" programming languages. They are too high level to be translated into machine language (with current technology). For this reason, they are interpreted--as is done with Java; hence they tend to execute relatively slowly.
Whether or not the world would be a better place with more extensive use of languages such as APL and LISP is moot. The point that should be made (I believe) is that (i) there is a lot to be learned from the experiences with APL and LISP and (ii) those experiences seem to be overlooked nowadays.
I could give many examples to illustrate this point. Most such, though, have already been stated many times in the academic literature. For example, J. Backus (referred to by Dr. Pollack), in his 1977 Turing Award lecture, outlined the advantages of applicative languages. And K. E. Iverson (inventor APL), in his 1979 Turing Award lecture, outlined how high level languages can be a "tool of thought", i.e. an aid to programmers' thinking. (Turing Award lectures are published annually in the Communications of the ACM.)
In fact, though, APL and LISP are not just programming languages. Each is a programming environment (which includes a language). Users of LISP, for example, have what appears to be a LISP machine--i.e. a virtual LISP computer. This machine does everything. From the users' perspective, there isn't a separate operating system. This means that the user only needs to learn about one thing: LISP.
The contrast with more common programming environments is large. For example, a C (or C++) programmer also typically learns some Unix. C and Unix, though, are entirely separate things. And they don't really mix (e.g. a C program that executes Unix commands). Which is why Perl was created. Which adds a third layer of complexity to the programming environment. Which raises the skill level required for programmers and the complexity of software.
I think that this is wrong--that the correct approach is to (re)design things so that the programming environment is "monolingual". Java might be considered as something in this direction.
Although APL and LISP have been around for decades, the recognition of the benefits provided by monolingual programming environments generally seems to be more recent. Some early work is by J. Heering & P. Klint, "Towards Monolingual Programming Environments", ACM TOPLAS, 7: 183-213 [1985].
I found it interesting that someone researching what might be tomorrow's most important and advanced technology appears almost a bit nostaligic for old largely-forgotten programming environments. Maybe he had a reason.
no forget it, just call me offtopic.
Copyrights are about coercing people th think that a given piece of information has the same value that you think it should have. It is morally no different than coercing beople to believe the same religion that you do. It's no wonder Hollywood has so many over-inflated ego's while people like Linus who has been over 100000 times more productive for society do not.
\begin{rant} Whether I like or not is not the matter, it is a fact that money is scarce, in limited supply, not everyone has the same amount,....
Given that a country that justify itself by using free speech to allow corruption by saying that money is speech is just that : a country trying to justify corruption.
I have any amount of speech I want, I can go out in the street and declare that the end is near, that Clinton sucks, or Gore or Bush soon, that strawberry topped beef has an horrible taste,... I don't have any limit to my capacity to generate speech.
Not so with $ (or £ or Euro, or lira,...). They are scarce, so not everybody can have the same amount so saying that it is speech means that you give more speech to the wealthy and make the not as much wealthy unequal in its ability to practice free speech. This is not a democratic system, and making it legal is one of the reasons why I want to laugh when somebody says that the USA are a democratic country.
\end{rant}
"The obvious mathematical breakthrough would be development of an easy way to factor large prime numbers." Bill Gates,
One could only be so lucky. Extinction of the human race-- I think it would be great. I think it would be beautiful. Beautiful as a mother gives birth to her child, raises it, but years after passes away. The child lives on, carrying out her virtues. Artificial is natural- a natural mechanism for survival. Chances are, AI will be able to survive much better in the future, than we will be. It will be for the better, IMHO.
Yeah, okay, I'm sounding like a mad scientist...
"I don't want the world, I just want your half"
(LISP is the language of symbolic programming, and though I'd rather do my stuff in C*, it does cater to AI programming.)
Yeah, there are only three books on my shelf. Hey, I'm still in my first year as a Master's student in AI. I'll get more books eventually ;)
-cryptomancer
Great at real and discrete math, and a natural born programming genius. No modesty.
Yes, we understand these tags always apply: fud, dupe, typo, slashdotted, topic name
As a young economist and philosopher (which gives you a clue which univ. I went to), I wrestled with this one for a long time, and came to the conclusion that Jordan is wrong on this one. There is just too much variety of human behaviour to ascribe it to seeking to maximise a single maximand. You end up with a conception of "utility" which is completely tautologous -- ie utility is "that which makes people do what they do"
For example, consider someone who hates himself and wants to frustrate all of his goals, but doesn't realise this fact (the fact that subconscious beliefs and motivations are possible creates yet more problems for utility theorists). What will this person do? Well, in fact, they tend to just become inactive, severely depressed people. In so far as this is a result of genuine subconscious desires (which is to say, in so much as a talking cure is possible (which is to say, not in most cases of clinical depression)), a talking cure for this kind of depression often involves pointing out to the person concerned that at some level, they are hanging around feeling depressed because they want to be. Then they stop doing it. If the cure isn't complete, they then realise that they have a new goal, attempt to frustrate themselves in this, forget the cure (subconsciously, they want to forget the cure) and sink back into depression. All the while remaining terribly unhappy. How can you explain this thoroughly predictable behaviour pattern by using utility?
Utility also has problems with anomie, religious experience and all manner of common human behaviours. As a heuristic for social science, it's excellent. But it shouldn't be elevated to a principle of logic or rationality.
-- the most controversial site on the Web