Thanks for the clarification. It's been a while since I read about the details of the network protocol stack, so that was a bit blurry in my memory.
I had the layers mixed up, and the Wikipedia article was helpful.
It sounds as though you're thinking of a caching HTTP proxy server, which isn't a router.
It doesn't have to understand XML to redirect a connection request for a URI.
I vaguely recall that when the web first came online, some routers/gateways were doing a lot of this, and
that web publishers complained because it was stealing potential ad revenue or
obstructing the ability of a host to change what it was publishing moment to moment... and, absent permission,
it was probably a copyright violation.
Anyone reading along who can tell me if I'm misremembering?
(I wasn't sure of a good search keyword to look up the history of this.)
Maybe such caching wasn't a bad idea--maybe it was just bad to do it without asking.
Another potential solution: Have browsers keep the DTDs cached,...
Or the routers. Frankly, if the result is known to not change, w3 could
probably agree with the network authorities to put copies around the net and
treat those heavily used URIs as URNs and just never got to w3 (or rarely go there)
instead.
The notion that URNs have to be known in advance as "the popular thing" rather than being discovered after-the-fact by noticing high-volume URIs is probably the real bug here.
Now that they've decided to abandon my code for another product that replaces its function, I'd like to continue working on my project as well as open it up to the world.
I'm not a lawyer but...
It sounds like they've paid you for this work. That probably means they are the copyright owner.
Copyright does not transfer by accident. So if you haven't got a document in hand by someone who is
a principal of the company that unambiguously identifies you as the new owner of the software, you
should tread very carefully.
Also, you're talking about having them 'gift' it to you, effectively. You apparently
did work they paid for and now you want to own it. That can't happen by magic, and most companies don't give away
assets. You might want to try 'buying' it since then there would be
a contract you could point to, and you would know who sold it. (I don't know if that makes it better. Ask a lawyer. It just seems to me like it might make a better paper trail.)
You don't say whether the other product is one your company is making or buying from outside. If the company
maintains a competing product, your non-compete agreement may be in play.
You might consider writing it again, clean, on your own time and machine, logging the intermediate versions
so they can be shown to be different than the backups the company has of its intermediate versions.
That may not be enough even. It might address copyright but not non-compete or trade secret.
You might consider getting the company to open source it instead of you. The difference (I think) would be that it would
be they, not you, who retains the right to make amended agreements with different conditions than the basic license.
In that case, all you need is that they open source it in a way that gives you the necessary rights of use, which may
be easier to establish than ownership. Also, in that case, you can probably get the company's lawyers involved in
making the license, and all you have to do is worry about whether you can use the license that is finally created.
In that case, you've evaded the worries about whether you transferred ownership right, and you're down to just "did they
pick a good license.
Did I mention I'm not a lawyer? You should not use this message as a guide to what you can do. Mostly you can take stuff like this that people like me write as conversation starters when you finally get serious and talk with someone who is legally competent to advise you properly.
And, by the way, if you make a mess of this and publish something you don't have the proper rights to,
you make a problem for people downstream in the user chain. There was a recent Slashdot article where something vaguely of this kind may have been in play. Even if not the same root cause, it illustrates a scenario you don't want to find yourself in.
This kind of approach has the hidden danger that once you fall into a certain
crowd, it's hard to dig your way out. It substantially increases the importance of choosing the
right one because you might never climb out.
Consider how many people think they are Democrats or Republicans just because
their parents are. (Parents are just an example, so don't be too quick to say
that parents aren't the chosen network. There will be some chosen network
and unless its attributes are freely advertised, you'll be signing up to have
things done for you in ways that are subtle and related to others you think you
know. It might just as well be "those drug fiends you kids run around with".)
Until the mid-1990's, I used to subscribe to paper magazines about technical topics.
And I'd get a lot of junk mail from vendors offering me stuff. Increasingly, I
found they talked about object-oriented programming and other topics I liked.
At first, I thought all my topics were winning the hearts and minds of people.
But after a while, I realized they had just pigeon-holed me as interested only
in those topics. What started off as a benefit they were offering me was now a kind
of Hell I had to live in... I'm sure there's some relevant Twilight Zone episode I
should be referencing here, but you get my point.
Freedom comes with choice. One reason that a lot of people don't like political primaries
is that it limits choice. If you can control the primary process (which has traditionally gotten
very little oversight--though this year probably got more than average), you have a great deal
of control of the election. People focus on the election as the thing that can be tampered with, and they make a polite fuss about who gets invited to this and that debate, about who takes this and that money, about the price of media, and so on. But it's those things, not a few hanging chads in the vote itself, that probably really sway the election. The damage is already done by time you reach the voting booth.
And what if everyone in the network is trusting everyone else, and no one is at the helm?
Or what if someone deviates from the network--is that weighted low as anomalous or high as important
that it wasn't statistically predicted and might signify something the group should peer at?
I don't see leaving these questions to a search engine... I think people should retain this right and
responsibility.
What's really scary about this isn't reputation, but the notion of having a set of legal tools
for telling people they can't publish things you don't like.
Reputation has been an issue for a long time, and people have
informally learned to manage it. Enough of us have grown up with this
that we can even teach our kids how not to make an idiot of themselves
online before they're ready.
But what's escalating of late is this:
legal bullying of anyone who says something you don't like
Most any truth can be converted into a value judgment.
A statement that global warming is being caused by x could be turned
into a slander suit against x. And from there we could get to no information
available online about x. And that means the Internet cannot do a main
thing many of us rely on it to do, which is to provide us with a way of doing
oversight on the world.
Omniscient search engines right now are seen as good things because they can
find information. That's good when you imagine that the truth is out there
and should be found. But when the purpose of finding it becomes to suppress
it, the next logical stage in the evolution of information warfare will be to
make the truth harder to find so that it can't easily be expunged. To pass
it along secret, trusted channels so that it's not available for target practice
by the rich and powerful when they feel threatened. I'm not sure I'm looking
forward to that.
Look at what happened with Political Correctness. It started out like such an
obviously good thing--that people shouldn't say bad things, and the world would
be better. But it hasn't played out thus.
You want a business idea? Create an organization that isn't based on removing controversial information but on creating new information that vets or refutes other information. The nice thing about that is you can create more than one such source, so you can have them duke it out in the marketplace. When you have a business based on suppression, you can't have a competing business based on non-suppression. The information is either out there or it's not.
Truth is often hard to prove; to assume all information is removable unless its
truth can be shown without a doubt is a pretty high bar to set.
There are more roads to a society that lives in fear of censorship than just
an overbearing government acting by fiat.
I'm not entirely negative on the idea of removing some "information" in cases
where it's hurtful or protected by privacy rules. But not everything is of
that kind. Free speech and privacy have a kind of yin/yang thing going, where a balance must be struck.
If I had two strands of magnets, arranged with random orders of polarity, identical strands would be able to stick together... Dissimilar strands... [if] you wiggled them,... [would] be more likely to come aparts.
Well, the article says:
Genes have the ability to recognise similarities in each other from a distance, without any proteins or other biological molecules aiding the process
And who knows if the writer of the article has a proper understanding or is being suitably precise.
However, by "at a distance" I must assume it means "with others intervening". So yes, one avenue is what
you're perhaps intimating, which is that maybe they come close and then fail to separate because they are similar before they are measured, rather gravitating as a result of being similar. Depending on the experiment, it wouldn't distinguish. But if they have determined that the things just know from a distance where to go, the explanation you suggest might not stand up.
Someone once told me a story about how oak trees separate themselves out. (I don't know if this story is true, but I don't tell it for its truth anyway, I tell it because it seems a good paradigm that could be true and useful.) They said that the leaves are acidic and that the acorns don't like to grow in acidic soil. So they drop acorns everywhere, but the ones that come up are not near the tree. So they don't grow too close together. This could be misinterpreted as "An oak tree sees another oak nearby and tries not to grow too close." and one could puzzle about how they "recognize each other". But if the recognition is just "wow, this is kind of acidic for me" then it could just as well stay clear of non-oak-trees and for no good reason. I wonder if perhaps there is some subtle effect on the electrical charge, on the chemical composition, or something else that is more low-grade than what they're looking for but creates a local gradient of plausibility such that other items nearby look for a metaphorical slope to slide down (a gradient that seems plausible). This wouldn't have to be a high reliability truth, just enough to bias the odds in a way that made it more likely to succeed. In computer science terms, all it needs is to emit a "hash bucket key" that can be read at a distance, but not necessarily an "equality test". It wouldn't have
to be biological molecules. It could just be some simpler chemical marker that they aren't testing for.
England's Home Office Scientific Development Branch... is currently investigating the feasibility of using lip-reading software as an additional tool for gathering information about criminals or for collecting evidence.
Would it be asking too much to have this worded as "gathering information about possible criminals"? (Or "suspected" or "alleged" would be ok.) The text quoted above, which is absent such an adjective, comes straight out of the article, and may or may not be how the Home Office
refers to it, but anyone engaged in public dialog on this matter (and preferrably those people when
doing their research) should strive to be meticulous on this point.
As soon as one loses that little bit of description, one is able to be much more cavalier about the loss of human privacy involved. It's one thing to rough up terrorists at the airport--who doesn't want that? But "possible terrorists" is just a synonym for "everyone". So when we say it's ok to rough up possible terrorists, we're saying it's ok to rough up anyone. And we can learn to think twice about that. Likewise, when we say it's ok to surveil the lip movements of "potential terrorists", we're saying it's ok to log everyone's private conversations. So let's be clear about that.
Saying we're just watching the lip movements of criminals isn't right. If we knew they were criminals, we would (for the most part) be arresting them. (Yes, yes, we might sometimes leave them on the street to lead us to their friends. But I don't think that's the only use that this technology will be put to.)
And how long until someone's lip movements are taken as a confession. Or as a justification for an otherwise-illegal search? The word "not" doesn't involve much movement of the lips. Lip-reading "I did not kill him." could easily look like "I did kill him." Will we be telling people that in order to stay clear of these things, we need to be more clear about our lip movements, just in case they're misconstrued?
Perhaps a stiff upper lip will give way evolutionarily to stiffening of both lips when talking, just as a form of personal protection. How sad. And worse if, as seems likely, dedicated criminals eventually learn the skill of not moving their lips while talking, and so that really only non-criminals become usefully tracked this way. Or perhaps it will become suspicious when one doesn't move one's lips, as it's probably inappropriately regarded by law enforcement as suspicious when one encrypts things. Then there will be the uncomfortable choice between hiding your communications and looking suspicious, or exposing your communications to misperception.
The data is out there. Lips convey meaning. So it's inevitable that this
technology will occur. But the uses to which it may reasonably be put are
in control of the people--at least in countries where the people have some
say in government. Let's hope they build up some reasonable guidelines on
appropriate vs inappropriate uses quickly.
Re:Anthropomorphizing obvious simulation result
on
Robots Learn To Lie
·
· Score: 1
When a human does X, you say it's "lying" but when a machine does the same thing, it's not?
Saying that one is not convinced lying is going on here is not saying that one doesn't think that
computers are capable of lying. I think the bar for showing an awareness at the level of what we humans call lying was not met, at least in the description I read. But that doesn't mean I think it couldn't be met under some other circumstances. To criticize this particular experiment is not to criticize the concept of robots as potential equals of people.
See your own thoughtful post on an unrelated topic for a reasonably lucid explanation of why you should tolerate more dissent in this discussion without construing the criticism in an overly broad and dismissive way.:)
funny you say the economics of special-purpose devices as opposed to general purpose robots won't work, because I think it's the other way round (at least for the medium-term). There is going to be a huge market for billions of cheap special-purpose robots like Roomba but for just about every imaginable task
Well, industry would prefer to sell it that way because then they can sell the same thing to you
over and over. Like selling you a vcr and a dvd player. Then a combined one. Then an hd player. Then combined hd and dvd. I suspect robots will undergo that. A kind of ongoing consolidation of selling you an X, then a Y, then an XY, an A, a B, an AB, then an AX and a BY for people who have been wishing for that combination, then finally an ABXY. It moves a lot of hardware when every year people re-buy the same thing just to get the latest and greatest. And in an age when intellectual property
is hard to protect, industry would like that.
But meanwhile we're probably knee deep in junk yards and stuff we're not recycling. So I hope it doesn't go quite that way.
If it does, we'll pollute the planet beyond all recovery. Probably we have already. Because no one
is making the commercial cost of these toys include their downstream disposal, and that's artificially
lowering the apparent price, creating some of the market euphoria that perpetuates the idea that the
planet can afford the trend.
that market is specifically NOT going to want unpredictable devices
The market demand will never be for that. It will be expressed differently. The demand will be
for reliable devices... but powerful ones. Money will be sitting there on the table waiting for
the first person willing to bend to the temptation, and there will be people saying "Aw, what can
it hurt. I'll take the money and hopefully it won't be THAT risky." They'll hire people who say
they can do it. They'll convince them it will be handled in testing, not dev. Then they'll convince
the testers that it's been funded and developed and there's nothing to be done but let it go or the
whole organization will fail. People will find reasons to send it out, so as not to be laid off.
Cognitive dissonance and all that.
It's also hard to predict what the impact will be of the inevitable development of machines more intelligent than us; that could bring anything from a complete Utopia to a complete dystopia.
Human nature being what it is, I'm not bullish on us surviving the next century or so of climate change, resource depletion, etc. Forget trusting robots, there's a lot of trust-in-people issues
to be overcome in the interim. Given the odds against us at this point, I'm almost willing to call it a definitional Utopia if intelligence in any sustainable
form, human or robot, survives what's to come. Alas.
Anyway, I appreciate the thouhtful conversation. I have disagreed at most every step, but I've enjoyed it. If you
have any parting thoughts, feel free to make them and I'll plan to read them, but probably not to
reply.
If your premise was to assume that in the future all robots are going to be highly intelligent, then sure, I agree with you, but I don't necessarily see that becoming the case, I think very few will be.
Nothing so grandiose. My point was rather that there are too aspects to intelligence (well, probably
many, but let's focus on two): one is having the "componentry" of intelligent behavior, that is, one-or-more general purpose problem solvers (whether or not any good) and a bunch of sensors and actuators; the other is "being able to use those devices well". When you say "highly intelligent", no, I don't mean it in the sense most people would think of that. But I mean it in the sense that a 2-year-old might be seen to be a great intelligence. It sees things, it runs around, it has a primitive goal structure and ability to reason, it has no common sense, and a great deal of ability to do a lot of damage. That I expect there to be a lot of.
An apparent premise of yours (though perhaps not one you intended?) is that we will build a number of special-purpose devices limited to a
single task, and I don't expect that to happen. The economics of it don't work. That's like seeing
the early game industry, and the free-standing Pong game, and assuming that future game processors
would contain exactly one game. Or like assuming that future calculators would do exactly and only
what was on the buttons. Everyone knows that programming is maleable, and there will be instant
market pressure for that. Besides, just as Star Trek has driven other things, everyone--even some
who have seen and shed a tear during The Measure of a Man (Star Trek: The Next Generation)--will want to own a Data. There probably won't be one worthy
of the character for quite some time. But there will be money to be made by selling anything that
can be plausibly claimed to be close. (Nor will the data be all in until it plausibly supports a claim of being "fully functional".)
You could try to convince me that people will have self-control and won't allow pushing the line
of what's possible--that they'll demand prudence in what capabilities are given such a device.
But, if you were so inclined to take that tack, you couldn't even
answer for email and web browsers. You'd think we as human beings would not have relinquished the right of the
machine to start programs on its own just because it or someone or something thought it was a good idea,
but that's precisely what happens when client-side scripting occurs, for example. It's the delivery
backbone along which viruses and worms travel. And the primary reason it's a threat is that human beings are too lazy to actually answer questions every time one of those things needs to run, and
not well enough qualified to answer the questions anyway. To compensate, they haven't gone to school
to learn more or allocated extra hours in the day to answer queries. Rather, they have simply lowered
the bar on safeguards they are willing to tolerate until things are able to run on their own without
human intervention, all in the service of saving time and energy. Why will robot safeguards follow
any different path?
If there's something to fear it's not robot intelligence. It's developmental phases, like the terrible twos, that we'd have to live
through to get there. And while I'm not bullish on the Singularity coming any time soon, I don't think the earlier,
pre-Singularity, phases are something we can ignore.
note that we have no evidence that there is a general purpose intent to lie, only a case where communication was used and observed to score better in one mode than another
1. a false statement made with deliberate intent to deceive... 2. something intended or serving to convey a false impression... 3. an inaccurate or false statement... Dictionary.com
There's more definitions, but this activity fits two of the top three (actually, at least four of the top seven) definitions of the word!
Since the first two of the definitions you cite require intent as part of the definition, and since
I was discussing "intent to lie", that's an odd claim. Except in the somewhat metaphorical use like
"That thermometer lies when it says it's 37 degrees out.", lying is an action of intent, so you must
demonstrate, not define away, the intent component.
And, note particularly, that intent to lie is not intent to have the end result. If I give you two
buttons and I say, "press one of these two buttons to do something (I won't tell you what) and then seek
the goal and score yourself, keeping track of which button to press", you may well learn to press the
buttons but that doesn't mean you know what you're causing, other than a higher score. Do we know the
program is modeling anything else besides its food score? Its own quality of life, for example? Much less the quality of life of its compatriots? Even if it thinks letting something go to its death is
going to give it a lower score, is that the extent of what it knows about death? Or has it seen
Bambi? I think the bar for lying is higher than you're setting it.
There are probably many animals
that know to do actions that have the effect of misleading, and yet we wouldn't say such animals are
liars unless they have the cognitive modeling capability to understand the difference.
Even the instructions to a child or to the impaired given in court to someone who will testify
(at least on the TV shows I've seen, which I must assume mirror reality) seem to focus on
the question of only allowing someone to testify as to the truth of something if they can model
what it is to lie. If they can't form the difference between truth and a lie, the information cannot be taken as an assertion of truth. It's just a data point.
There's an excellent illustration of some of these issues in the book Ender's Game. Orson Scott Card explores a lot of strategies and meta-strategies in there, and there's a notable one at one point where this comes into play clearly. I don't want to spoil anything by saying more about that book, which is a fun read. You'll know it when you see it.
This was specifically basically an 'evolution simulation', so by design these robots were able to develop original new behaviours to their individual (or possibly 'species') benefit (it's not some inherent emergent behaviour or quality of even neural nets - this required significant additional effort that can't really come about by accident). Generalising this behaviour to 'all robots' is not correct...
That "really can't come about by accident"? I don't buy that.
I'll give you that it may well have been not by the "conventional" means. In fact, it's very hard
to read useful data into a news report that merely describes the behavior of the bots, and doesn't
say why they took the behavior, and especially a news report that uses phrases like "eerily wicked",
"calmly", etc. to describe behaviors. But ok, let's give you at least that the evolutionary form
had some effect.
Note well that in evolution, not every aspect of the machine evolves. Some parts are held constant.
So an evolved model of something may have a learned behavior that is learned not because of the evolution but merely because of the situational structure and the basic learning mechanism.
But, as I said, let's ignore that and assume you're right on this, because it will make my real point
simpler.
Even if we knew evolution was used as the catalyzing mechanism, I think it would be an
extraordinary claim if you were able to say that this is something that we were safe from
seeing in a non-evolved robot that had the basic ability to evaluate and score its environment,
select its goals, etc. I assume someone has done the (probably not difficult) legwork to show
that this already implements a Turing machine. (I assume that mostly on the basis that a Turing
machine isn't very complicated, and these robots sound moreso... I'd allege
a good strong stake in the ground like "eerily wicked" implies Turing powerful. Though maybe
someone is going to tell me that one of the evolutionary steps was provably to cross the Turing
power line... I'm guessing it started already well across that line or it couldn't have performed
at all, much less evolved, in the first iteration and/or it couldn't have reached this level by
only 50 iterations. But it would depend even then on the initial
construction and I'm willing to be surprised.)
But my point is that once Turing power is reached, and clearly we won't be stopping robots from doing that, I don't see how you can speak
comfortably that this behavior won't happen due to regular neural nets. All you can say is that it
hasn't. But I don't see anything fundamental about the computations involved in evolution that are
likely to be any more elaborate than running a hypothetical, seeing the results, adjusting one's
parameters, and trying again (vaguely like happens to the computer at the end of Wargames if you can stomach a somewhat whimsy visual metaphor in a serious discussion of computational complexity).
People learn and re-learn processes, and I expect us to put that into robots. And even if we don't,
the word "emergent" suggests that if it has survival value, it will end up happening.
It seems to me that it would be an extraordinary claim if you could really prove your
statement
it remains the case that a more deterministically programmed robot is highly unlikely to ever develop such unpredictable and undesired behaviour.
That would seem to me tantamount to a claim that there are two different kinds of intelligent processors
and that you know a class of computations that can only be done on one and not on the other.
To my knowledge, the ability to seriously prove such a claim is the holy grail of much investigation
in computation science, and many hypothesize that it simply cannot be. But whether it can or cannot
be, I don't think it's known now.
This is HIGHLY disturbing. Even if this is just a fluke or a bug, it shows what can happen if we give too much power to robots.
While this kind of stuff creeps me out as much as the next guy, and while it argues for being
careful about what we trust robots to do, it's something we should know anyway because there
many ways our trust can be violated without a robot lying. By far the more likely way they're going
to let us down is just to exercise poor judgment. That is, to search for something that looks like
a peanut butter sandwich but is really a rag with some grease on it... Getting the small details of
common sense wrong is just as dangerous as anything deliberate.
What we really learn here is that the mathematics of learn things like lying as a strategy
isn't remarkably
complex; that is,
(that is, the number of
computational steps required to discover it works in at least some cases is small... note that
we have no evidence that there is a general purpose intent to lie, only a case where communication
was used and observed to score better in one mode than another). This is not a story about robots, it's a
cautionary tale about neural nets, what they measure, how they fail, etc... and we didn't invent the idea of neural nets--we found
it already installed in every living thing around us.
I went to the Museum of Science in Boston a few months back and saw, in the butterfly exhibit,
a moth that had evolved coloration that was indistinguishable from an owl's face,
hoping to scare off predators that were afraid of owls. Probably that's the more sophisticated result of the same notions.
And yet it occurs in an animal that isn't, as a general purpose matter, a very sophisticated
animal. Most people would find already-extant robots more socially engaging
than a moth. For example, a moth is not capable of even serving up a beer during the game or vacuuming
up the mess after your buddies go home.
So take heart: The likely truth is that this is unavoidable.
If all it does is teach us to have a healthy skepticism for unrestrained technology,
it's actually a good thing. We needed that skepticism anyway.
Emergent behaviour is tricky and unpredictable pre-facto.... we can expect emergent behavior that is as different from that of computers as a clock's is from the cogs that form it.[1]
I agree with your remarks regarding emergent behavior and I liked the clock analogy, which I assume was
in some ways a variant of Searle's Chinese Room.
As you point out, trusting in some kind of "control" over that system is rather silly, at best. I believe that it is in fact both undesirable and impossible.
I agree, too, that it's impossible to control. However, I think it's useful to protect oneself from it,
just as it's useful to place limits on government.
Although it's also true that the term "it" is a little elusive here. Because, for example, it's easy to
indulge the illusion that the "it" is the Internet, and that if you flick a switch, you're free of it.
But the "it" is the "connected society", and it remains connected even if you turn off individual pieces
of it. For example, whether you have the Internet or not, you can still be affected by identity theft,
by insurance companies discrimintating based on genetic data, by someone in another country reading how
to create an atomic bomb and then creating one, etc. Some of these were risks even before the Internet,
when there were only paper libraries and paper telephone directories. The Internet has both accelerated
that, and created combinatorial effects that are hard to quantify.
It's this full range of things I mean
to attach when I say there should be protections. Not against data on a wire, but against the use of
information. It can't be stopped, but it should be confronted as the real effect it is, and not ignored
as if we must simply endure it.
Internet still has no center, technically speaking, control can now be wielded, through software code, from anywhere. What's different, in comparison to the physical world, is that acts of control become harder to detect and those wielding control more difficult to discern.
Or from nowhere. The risk of a bad guy taking over is serious, but the risk that no one is at the helm is
much more likely to lead us to death by Global Warming, for example.
You have to look no further than the US Congress to see a worked example.
If you idealize every single member of Congress as intelligent,
and I think a similar analogy can be made for people on the net or for companies on the net
(where you still have to question intelligence sometimes, but let's not and say we did),
it's pretty clear that the problem isn't just the sinister taking hold of someone with total power.
It's also that it's easy to cause behavior that no one can take responsibility for, and that isn't
in the best interest of individuals. The
Internet is no different, but not because we didn't have examples of this before.
Just because we didn't heed them.
A free tool that is able to read but not write (and hence would not equal giving away the entire program) could be made by, to be slightly facetious, commenting out the lines of code that enable buttons to be pressed and puts icons on the tool bar.
Well, it would depend on whose definition of "free" you were using. This would satisfy your or my definition
of free, but would have the problem that, absent source code, it wasn't GPL free. And if it came with source
code, then putting back those lines and recompiling to get the full application wouldn't be hard.
So anyone trying to protect their proprietary investment by just doing simple editing like this might not
really be doing so.
while your view appears to be that software copyright should not exist, mine is that it should, so on that we differ.
It's not very relevant to this thread, but since you mention it:
I am a strong advocate of the importance of copyright per se, for both programs and literature.
Where I might have confused you is that I don't think copyrights on
software published software that is 10-20 years old are really that valuable, and I think
that society should change the duration of software copyright downward. The
commercial shelf-life of software just isn't that long. I
think a much greater public good is served by opening things up the possibility of reuse at
that point. Books, by contrast, have a much longer commercial shelf-life--a well-written
novel might sell for many years. So I can see a longer period justified based on the
nature of the work that is coprighted. There are other objective differences between
software and literature that I could cite as well, but this isn't the forum. Bottom line,
though, I do believe and endorse the statement "copyright should exist".
The strongest argument for not reducing the duration of software copyright is that
to the extent that old software might be resold without being first brought up to date,
it risks re-introducing a lot of buggy and unsecure programs back into the world.
In that regard, vendors may actually do a service by withholding such old and probably low-value programs
out of
some more personal motive.
it would be sufficient for the sake of old documents to provide a free tool that is able to read those documents, or a tool that would convert them to an open document format. This tool wouldn't need to have its source published.
As noted in another post about
this article, it may be that there is no "format" other than "the code". If so, then the only free tool
that is cheap to make is a wrapper around a complete application that just calls only part of that application.
If so, making the wrapped tool free means giving away the entire program, not just the file part.
In effect, then, this amounts to requesting that old versions be made free. Any difference between your
proposal and asking old versions of their editor to just be made "free" (for whatever free you might be
meaning) is just words, I suspect--nothing semantic.
Of course, this comes back to the question of whether there should be software patents at all, and whether
software copyrights should have the immensely long durations that they do. Indeed, at some point, probably much shorter than happens now, having
old tools be free so they can be recycled for other purposes may not be bad. It might even give vendors a
kick in the pants to move faster to make newer tools be different enough that the old tools didn't threaten
them. But bypassing a proper change in software copyright and patent law and instead just beating up on certain
people who have things one wants does not seem the best approach to me.
If the software is useful and it was generated by a one man band, where else are you going to get the better support?
If the application is useful, well-written, and well-documented, why would it matter who you went to?
The problem with giving away source code is that this answer is far from obvious. What's most scary about your answer, in fact, is not the fact that you suggest this as a possible outcome--certainly it's possible--but that you strongly hint that there's no other possible answer.
To start with, companies that bet their operations in software will not be using this particular application, so a lot of what you are saying does not apply at all.
Again, you are asserting this as if there is no alternative. And yet, again, with no pointer to hard data.
As if somehow saying it were so will make it so.
But smaller companies (other one man bands perhaps?) may find this useful and will pay for the support.
To what degree of reliability? Based on what data? What if you're wrong? How would you rate your chances of being wrong?
3. I'm more familiar with Java and Javascript, but I presume that Flash uses a similar sandbox, restricting what Flash applications can do. For instance, Java applets downloaded from example.com can only make network connections back to example.com, and cannot read or write local files etc. 4. If it is anything like the Java applet sandbox, then it shouldn't be able to access general webpages, only the page it was loaded from. WPA and the administration password for your router have nothing to do with UPnP IGD.
Ah, ok. If it could access arbitrary pages, I was worried that it could access the admin component of my
wireless router. (From the outside, WPA protects me from people accessing that admin stuff, but from the inside, it does not... only a more vanilla passwording protects me inside.) But if it's something like a limited access to the net, and if that limitation is not vulernable (nothing ever seems as safe as one expects), then it sounds like I'm fine.
I certainly do not have UPnP enabled now. My Belkin router fortunately comes with it off though offers the following remark in the online help, which given this recent news story I now see has a double-edged interpretation: “An application that is UPnP compliant has the ability to communicate with the Router, basically "telling" the Router which way it needs the firewall configured.”
I had just worried that some worm could find my gateway and enable it so that its friends could come in... but I read what you're saying as saying that if it got as far as being able to do that, it wouldn't need such an entry point anyway, it would already be inside and set up for business.
One follow-up question on this matter of only having access to the web site the item came from. A long time ago, I looked at security models in an informal way, but I am not an expert in this, so don't take my remarks about what I thought as any kind of authority--just as framing the question: I had thought that there was a subtle distinction between an applet on a page and a downloaded plug-in, and that it hinged on the question of whether the code was trusted. If Flash is a plug-in, not an applet, and is therefore trusted, and Flash is the thing that is doing the network connecting, then is Flash the thing that is holding my security in its hands, or does it have some way of making sure that the program it runs is subject to a more restrictive security model, such as I'm imagining applets to have? If I've asked the question in a way that makes it hard to answer because I've presupposed some division that's not there, feel free to repair the question before answering it.
Perhaps I should go read a recent reference on network security. But, sadly, there will be many with
questions like these who don't have the time or training to do that, so hopefully the question lends itself
to at least some sort of minimal high-level summary that doesn't require that.
The vulnerability is really Flash not restricting what untrusted scripts can do. [...]
UPnP does expand the possibilities for a malicious Flash app to initiate connections with your machine. But unless Flash also allows you to open server sockets, the attacker would also need to find an exploitable service running on your machine.
Excuse my ignorance/confusion, but... I'm not up on the details of either Flash or UPnP, and yet I still need to understand this better and so I have a few questions.
Is the Flash being discussed the Flash player for a browser, right? (Not some sort of Flash related to flash memory and the BIOS and/or USB Flash drives? And the Flash issue is not in the router?)
Why is there a difference between the Flash vulnerability in different browsers? What's the basis of
the protection? Is it because the
player binaries differ between browsers, or because the security model of the browsers differ, or what?
If a Flash player is running malware already, why does it care any longer about the router? Isn't it
already in my machine, and hence inside my network? And can't it generally get out quite easily with
whatever data it finds without further problem? Or is there some security model limiting the actions
of the Flash player to only certain operations?
Is it forbidden from writing files, particularly executable files? I assume a virus utility would
notice this, but maybe since it's a trusted plug-in, it wouldn't?
If it can access
web pages, isn't there also a potential vulnerability that many routers are configurable from inside the
firewall over the network? In that case, couldn't it reenable UPnP itself? (Even if it was forbidden to read files from the disk and access the net, couldn't it just do the web page modification and then wait for a later copy of itself to arrive on a separate occasion to exploit the previously and silently opened hole?)
If that can happen at all, will having a decent password
for one's firewall reduce this risk? (Even though I have WPA-PSK enabled and a pretty long password, internal connections to a router over a secure connection seem like they're going to succeed because of the PSK, leaving the router's admin password the only thing in the way... or is there some other fortunate barrier?) Do routers tend to protect themselves from internal exhaustive or dictionary attacks? Would a virus protection tool notice this, or would it just think it normal that a browser was opening lots of web pages? In other words, do I need to switch my router to be configurable only over a serial link? (Even if I did, would I be vulnerable while the serial line was connected?)
If there's just a FAQ with answers to questions like these, please point me to it.
I read the article, but it was pretty thick with device and protocol and program-specific jargon that even a technical person might not understand,
depending on their areas of expertise.
Since it's open source there is no real point in charging for the software, but you could charge for support.
I wouldn't bet my family's ability to eat on this.
The problem isn't that there isn't money to be made on support.
The problem is that a single person working his garage on software is not
a support organization. Consequently, when it comes time to release that software,
serious people (the kind with money) with a serious need for support (the kind they're going to bank
their business on) won't trust a single individual for support.
So your software will flow freely out, the support need may arise and support dollars may exist, but
it's quite a gamble to assume they'll flow to you rather than to someone with a brand name
and a committed resource of people that can stand behind a claim of support.
It's easy to want to believe you're going to get the money.
But you are loads safer and much more likely to be right if you assume someone else will get it.
You'd better assume you're going to get nothing and be satisfied with that.
People may tell you otherwise, but I'd be surprised if they'd place money on that bet.
Advice, especially from evangelists, is cheap to offer when the one doing the offering doesn't have to deal with the consequences of being wrong. Don't let anyone convince you that doing something that is in the best interest of you and your own financial needs is some form of paranoia.
Of course, it's possible that the willingness of others to make free software has sufficiently
driven down the price of software that there's no money you can make by selling it either.
That's a different matter entirely. In that case, maybe you can sell service (supported by your own
software, without releasing it), or maybe you just have something the world regards as worthless.
That would be sad. But being sad would not make it impossible.
Everything else is free too, if you are willing to ignore the law, which a lot of people are.
Part of the problem is: how does someone know it's free? Certainly the fact that it comes affixed with such a label doesn't tell you that--how do you know you're getting it from someone who is reliable if it comes in a free pack from anyone. If a van pulls up on the street and offers you something that looks like a consumer item at radically reduced price, you're already on alert... is the stuff stolen?
You can get a lot of nearly free versions of DVDs of movies off of the net, too, but they turn out to be bootlegs.
If people are charging media cost for disks of so-called free software, software that obviously took time to produce, how is that not going to look the same to someone who isn't technically savvy? Sure, to someone
who is comfortble browsing Sourceforge it might work, but to that person's mom?
The trust issue is a serious one, and the open source community deals with that by a set of mechanisms that involve finding trusted sources. But just handing someone a disk and saying "you can use this" is not it. It requires technical and legal sophistication to know you're not becoming (or just paying) a criminal by buying. It also requires fair technical sophistication to know where to get updates, etc...
You're not necessarily doing a non-technical person a favor by getting them into this. It's like getting someone to buy a car that has to be maintained oneself instead of getting them to buy a name brand with people that can maintain it. That's great for people who like being tinkerers, but not for everyone. And I'd think most people who can't fix their own car would be pretty suspicious of buying from folks who give away cars for free and said "don't worry about the cost, I'll make my money on the maintenance."
Are we talking per-customer watermarks? (The article didn't seem to say.) Aside from the usual privacy
implications, that would have its own problems,
since it would allow for unbounded downstream prosecution of anyone who ever let even one copy go free, including
through malware. It would make it quite a liability to even buy such stuff.
I have to wonder why everyone's trying to put a band-aid on the problem.
Because politics rarely involves one person getting their way, no matter how good.
So the practical person, for better or worse, examines compromises.
Here's a wacky idea--build a nuclear plant or two, and provide the energy that people are demanding.
I'm actually seriously with you on that.
Lots of people fear nuclear. But if you really believe either he global warming issue, or the
fact that we're peaking on oil, we're going to need cheap, clean energy soon.
I'm all for wind and solar, but don't expect that to deploy in enough time to really work.
So your band-aid (heh--see, everyone has one--it's only natural) of filling the gap with a bunch of nuclear plants sounds like the only
way forward that makes much sense to address the cause, not the symptom... at least in the next 10-20 years.
Just, please, let's put them above the plain that might be flooded by global warming. And definitely not
in the basement of anyone with a government-controlled thermostat.
Slashdot Mirror
Thanks for the clarification. It's been a while since I read about the details of the network protocol stack, so that was a bit blurry in my memory. I had the layers mixed up, and the Wikipedia article was helpful.
Indeed. Again, thanks for the correction.
It doesn't have to understand XML to redirect a connection request for a URI.
I vaguely recall that when the web first came online, some routers/gateways were doing a lot of this, and that web publishers complained because it was stealing potential ad revenue or obstructing the ability of a host to change what it was publishing moment to moment... and, absent permission, it was probably a copyright violation. Anyone reading along who can tell me if I'm misremembering? (I wasn't sure of a good search keyword to look up the history of this.)
Maybe such caching wasn't a bad idea--maybe it was just bad to do it without asking.
Or the routers. Frankly, if the result is known to not change, w3 could probably agree with the network authorities to put copies around the net and treat those heavily used URIs as URNs and just never got to w3 (or rarely go there) instead.
The notion that URNs have to be known in advance as "the popular thing" rather than being discovered after-the-fact by noticing high-volume URIs is probably the real bug here.
I'm not a lawyer but...
It sounds like they've paid you for this work. That probably means they are the copyright owner. Copyright does not transfer by accident. So if you haven't got a document in hand by someone who is a principal of the company that unambiguously identifies you as the new owner of the software, you should tread very carefully.
Also, you're talking about having them 'gift' it to you, effectively. You apparently did work they paid for and now you want to own it. That can't happen by magic, and most companies don't give away assets. You might want to try 'buying' it since then there would be a contract you could point to, and you would know who sold it. (I don't know if that makes it better. Ask a lawyer. It just seems to me like it might make a better paper trail.)
You don't say whether the other product is one your company is making or buying from outside. If the company maintains a competing product, your non-compete agreement may be in play.
You might consider writing it again, clean, on your own time and machine, logging the intermediate versions so they can be shown to be different than the backups the company has of its intermediate versions. That may not be enough even. It might address copyright but not non-compete or trade secret.
You might consider getting the company to open source it instead of you. The difference (I think) would be that it would be they, not you, who retains the right to make amended agreements with different conditions than the basic license. In that case, all you need is that they open source it in a way that gives you the necessary rights of use, which may be easier to establish than ownership. Also, in that case, you can probably get the company's lawyers involved in making the license, and all you have to do is worry about whether you can use the license that is finally created. In that case, you've evaded the worries about whether you transferred ownership right, and you're down to just "did they pick a good license.
Did I mention I'm not a lawyer? You should not use this message as a guide to what you can do. Mostly you can take stuff like this that people like me write as conversation starters when you finally get serious and talk with someone who is legally competent to advise you properly.
And, by the way, if you make a mess of this and publish something you don't have the proper rights to, you make a problem for people downstream in the user chain. There was a recent Slashdot article where something vaguely of this kind may have been in play. Even if not the same root cause, it illustrates a scenario you don't want to find yourself in.
This kind of approach has the hidden danger that once you fall into a certain crowd, it's hard to dig your way out. It substantially increases the importance of choosing the right one because you might never climb out.
Consider how many people think they are Democrats or Republicans just because their parents are. (Parents are just an example, so don't be too quick to say that parents aren't the chosen network. There will be some chosen network and unless its attributes are freely advertised, you'll be signing up to have things done for you in ways that are subtle and related to others you think you know. It might just as well be "those drug fiends you kids run around with".)
Until the mid-1990's, I used to subscribe to paper magazines about technical topics. And I'd get a lot of junk mail from vendors offering me stuff. Increasingly, I found they talked about object-oriented programming and other topics I liked. At first, I thought all my topics were winning the hearts and minds of people. But after a while, I realized they had just pigeon-holed me as interested only in those topics. What started off as a benefit they were offering me was now a kind of Hell I had to live in... I'm sure there's some relevant Twilight Zone episode I should be referencing here, but you get my point.
Freedom comes with choice. One reason that a lot of people don't like political primaries is that it limits choice. If you can control the primary process (which has traditionally gotten very little oversight--though this year probably got more than average), you have a great deal of control of the election. People focus on the election as the thing that can be tampered with, and they make a polite fuss about who gets invited to this and that debate, about who takes this and that money, about the price of media, and so on. But it's those things, not a few hanging chads in the vote itself, that probably really sway the election. The damage is already done by time you reach the voting booth.
And what if everyone in the network is trusting everyone else, and no one is at the helm? Or what if someone deviates from the network--is that weighted low as anomalous or high as important that it wasn't statistically predicted and might signify something the group should peer at? I don't see leaving these questions to a search engine... I think people should retain this right and responsibility.
What's really scary about this isn't reputation, but the notion of having a set of legal tools for telling people they can't publish things you don't like. Reputation has been an issue for a long time, and people have informally learned to manage it. Enough of us have grown up with this that we can even teach our kids how not to make an idiot of themselves online before they're ready.
But what's escalating of late is this:
Most any truth can be converted into a value judgment. A statement that global warming is being caused by x could be turned into a slander suit against x. And from there we could get to no information available online about x. And that means the Internet cannot do a main thing many of us rely on it to do, which is to provide us with a way of doing oversight on the world.
Omniscient search engines right now are seen as good things because they can find information. That's good when you imagine that the truth is out there and should be found. But when the purpose of finding it becomes to suppress it, the next logical stage in the evolution of information warfare will be to make the truth harder to find so that it can't easily be expunged. To pass it along secret, trusted channels so that it's not available for target practice by the rich and powerful when they feel threatened. I'm not sure I'm looking forward to that.
Look at what happened with Political Correctness. It started out like such an obviously good thing--that people shouldn't say bad things, and the world would be better. But it hasn't played out thus.
You want a business idea? Create an organization that isn't based on removing controversial information but on creating new information that vets or refutes other information. The nice thing about that is you can create more than one such source, so you can have them duke it out in the marketplace. When you have a business based on suppression, you can't have a competing business based on non-suppression. The information is either out there or it's not. Truth is often hard to prove; to assume all information is removable unless its truth can be shown without a doubt is a pretty high bar to set. There are more roads to a society that lives in fear of censorship than just an overbearing government acting by fiat.
I'm not entirely negative on the idea of removing some "information" in cases where it's hurtful or protected by privacy rules. But not everything is of that kind. Free speech and privacy have a kind of yin/yang thing going, where a balance must be struck.
Well, the article says:
And who knows if the writer of the article has a proper understanding or is being suitably precise. However, by "at a distance" I must assume it means "with others intervening". So yes, one avenue is what you're perhaps intimating, which is that maybe they come close and then fail to separate because they are similar before they are measured, rather gravitating as a result of being similar. Depending on the experiment, it wouldn't distinguish. But if they have determined that the things just know from a distance where to go, the explanation you suggest might not stand up.
Someone once told me a story about how oak trees separate themselves out. (I don't know if this story is true, but I don't tell it for its truth anyway, I tell it because it seems a good paradigm that could be true and useful.) They said that the leaves are acidic and that the acorns don't like to grow in acidic soil. So they drop acorns everywhere, but the ones that come up are not near the tree. So they don't grow too close together. This could be misinterpreted as "An oak tree sees another oak nearby and tries not to grow too close." and one could puzzle about how they "recognize each other". But if the recognition is just "wow, this is kind of acidic for me" then it could just as well stay clear of non-oak-trees and for no good reason. I wonder if perhaps there is some subtle effect on the electrical charge, on the chemical composition, or something else that is more low-grade than what they're looking for but creates a local gradient of plausibility such that other items nearby look for a metaphorical slope to slide down (a gradient that seems plausible). This wouldn't have to be a high reliability truth, just enough to bias the odds in a way that made it more likely to succeed. In computer science terms, all it needs is to emit a "hash bucket key" that can be read at a distance, but not necessarily an "equality test". It wouldn't have to be biological molecules. It could just be some simpler chemical marker that they aren't testing for.
Or not. I'm neither a chemist nor a biologist.
Would it be asking too much to have this worded as "gathering information about possible criminals"? (Or "suspected" or "alleged" would be ok.) The text quoted above, which is absent such an adjective, comes straight out of the article, and may or may not be how the Home Office refers to it, but anyone engaged in public dialog on this matter (and preferrably those people when doing their research) should strive to be meticulous on this point.
As soon as one loses that little bit of description, one is able to be much more cavalier about the loss of human privacy involved. It's one thing to rough up terrorists at the airport--who doesn't want that? But "possible terrorists" is just a synonym for "everyone". So when we say it's ok to rough up possible terrorists, we're saying it's ok to rough up anyone. And we can learn to think twice about that. Likewise, when we say it's ok to surveil the lip movements of "potential terrorists", we're saying it's ok to log everyone's private conversations. So let's be clear about that.
Saying we're just watching the lip movements of criminals isn't right. If we knew they were criminals, we would (for the most part) be arresting them. (Yes, yes, we might sometimes leave them on the street to lead us to their friends. But I don't think that's the only use that this technology will be put to.)
And how long until someone's lip movements are taken as a confession. Or as a justification for an otherwise-illegal search? The word "not" doesn't involve much movement of the lips. Lip-reading "I did not kill him." could easily look like "I did kill him." Will we be telling people that in order to stay clear of these things, we need to be more clear about our lip movements, just in case they're misconstrued?
Perhaps a stiff upper lip will give way evolutionarily to stiffening of both lips when talking, just as a form of personal protection. How sad. And worse if, as seems likely, dedicated criminals eventually learn the skill of not moving their lips while talking, and so that really only non-criminals become usefully tracked this way. Or perhaps it will become suspicious when one doesn't move one's lips, as it's probably inappropriately regarded by law enforcement as suspicious when one encrypts things. Then there will be the uncomfortable choice between hiding your communications and looking suspicious, or exposing your communications to misperception.
The data is out there. Lips convey meaning. So it's inevitable that this technology will occur. But the uses to which it may reasonably be put are in control of the people--at least in countries where the people have some say in government. Let's hope they build up some reasonable guidelines on appropriate vs inappropriate uses quickly.
Saying that one is not convinced lying is going on here is not saying that one doesn't think that computers are capable of lying. I think the bar for showing an awareness at the level of what we humans call lying was not met, at least in the description I read. But that doesn't mean I think it couldn't be met under some other circumstances. To criticize this particular experiment is not to criticize the concept of robots as potential equals of people.
See your own thoughtful post on an unrelated topic for a reasonably lucid explanation of why you should tolerate more dissent in this discussion without construing the criticism in an overly broad and dismissive way. :)
Well, industry would prefer to sell it that way because then they can sell the same thing to you over and over. Like selling you a vcr and a dvd player. Then a combined one. Then an hd player. Then combined hd and dvd. I suspect robots will undergo that. A kind of ongoing consolidation of selling you an X, then a Y, then an XY, an A, a B, an AB, then an AX and a BY for people who have been wishing for that combination, then finally an ABXY. It moves a lot of hardware when every year people re-buy the same thing just to get the latest and greatest. And in an age when intellectual property is hard to protect, industry would like that.
But meanwhile we're probably knee deep in junk yards and stuff we're not recycling. So I hope it doesn't go quite that way. If it does, we'll pollute the planet beyond all recovery. Probably we have already. Because no one is making the commercial cost of these toys include their downstream disposal, and that's artificially lowering the apparent price, creating some of the market euphoria that perpetuates the idea that the planet can afford the trend.
The market demand will never be for that. It will be expressed differently. The demand will be for reliable devices... but powerful ones. Money will be sitting there on the table waiting for the first person willing to bend to the temptation, and there will be people saying "Aw, what can it hurt. I'll take the money and hopefully it won't be THAT risky." They'll hire people who say they can do it. They'll convince them it will be handled in testing, not dev. Then they'll convince the testers that it's been funded and developed and there's nothing to be done but let it go or the whole organization will fail. People will find reasons to send it out, so as not to be laid off. Cognitive dissonance and all that.
Human nature being what it is, I'm not bullish on us surviving the next century or so of climate change, resource depletion, etc. Forget trusting robots, there's a lot of trust-in-people issues to be overcome in the interim. Given the odds against us at this point, I'm almost willing to call it a definitional Utopia if intelligence in any sustainable form, human or robot, survives what's to come. Alas.
Anyway, I appreciate the thouhtful conversation. I have disagreed at most every step, but I've enjoyed it. If you have any parting thoughts, feel free to make them and I'll plan to read them, but probably not to reply.
Nothing so grandiose. My point was rather that there are too aspects to intelligence (well, probably many, but let's focus on two): one is having the "componentry" of intelligent behavior, that is, one-or-more general purpose problem solvers (whether or not any good) and a bunch of sensors and actuators; the other is "being able to use those devices well". When you say "highly intelligent", no, I don't mean it in the sense most people would think of that. But I mean it in the sense that a 2-year-old might be seen to be a great intelligence. It sees things, it runs around, it has a primitive goal structure and ability to reason, it has no common sense, and a great deal of ability to do a lot of damage. That I expect there to be a lot of.
An apparent premise of yours (though perhaps not one you intended?) is that we will build a number of special-purpose devices limited to a single task, and I don't expect that to happen. The economics of it don't work. That's like seeing the early game industry, and the free-standing Pong game, and assuming that future game processors would contain exactly one game. Or like assuming that future calculators would do exactly and only what was on the buttons. Everyone knows that programming is maleable, and there will be instant market pressure for that. Besides, just as Star Trek has driven other things, everyone--even some who have seen and shed a tear during The Measure of a Man (Star Trek: The Next Generation)--will want to own a Data. There probably won't be one worthy of the character for quite some time. But there will be money to be made by selling anything that can be plausibly claimed to be close. (Nor will the data be all in until it plausibly supports a claim of being "fully functional".)
You could try to convince me that people will have self-control and won't allow pushing the line of what's possible--that they'll demand prudence in what capabilities are given such a device. But, if you were so inclined to take that tack, you couldn't even answer for email and web browsers. You'd think we as human beings would not have relinquished the right of the machine to start programs on its own just because it or someone or something thought it was a good idea, but that's precisely what happens when client-side scripting occurs, for example. It's the delivery backbone along which viruses and worms travel. And the primary reason it's a threat is that human beings are too lazy to actually answer questions every time one of those things needs to run, and not well enough qualified to answer the questions anyway. To compensate, they haven't gone to school to learn more or allocated extra hours in the day to answer queries. Rather, they have simply lowered the bar on safeguards they are willing to tolerate until things are able to run on their own without human intervention, all in the service of saving time and energy. Why will robot safeguards follow any different path?
If there's something to fear it's not robot intelligence. It's developmental phases, like the terrible twos, that we'd have to live through to get there. And while I'm not bullish on the Singularity coming any time soon, I don't think the earlier, pre-Singularity, phases are something we can ignore.
Since the first two of the definitions you cite require intent as part of the definition, and since I was discussing "intent to lie", that's an odd claim. Except in the somewhat metaphorical use like "That thermometer lies when it says it's 37 degrees out.", lying is an action of intent, so you must demonstrate, not define away, the intent component.
And, note particularly, that intent to lie is not intent to have the end result. If I give you two buttons and I say, "press one of these two buttons to do something (I won't tell you what) and then seek the goal and score yourself, keeping track of which button to press", you may well learn to press the buttons but that doesn't mean you know what you're causing, other than a higher score. Do we know the program is modeling anything else besides its food score? Its own quality of life, for example? Much less the quality of life of its compatriots? Even if it thinks letting something go to its death is going to give it a lower score, is that the extent of what it knows about death? Or has it seen Bambi? I think the bar for lying is higher than you're setting it.
There are probably many animals that know to do actions that have the effect of misleading, and yet we wouldn't say such animals are liars unless they have the cognitive modeling capability to understand the difference. Even the instructions to a child or to the impaired given in court to someone who will testify (at least on the TV shows I've seen, which I must assume mirror reality) seem to focus on the question of only allowing someone to testify as to the truth of something if they can model what it is to lie. If they can't form the difference between truth and a lie, the information cannot be taken as an assertion of truth. It's just a data point.
There's an excellent illustration of some of these issues in the book Ender's Game. Orson Scott Card explores a lot of strategies and meta-strategies in there, and there's a notable one at one point where this comes into play clearly. I don't want to spoil anything by saying more about that book, which is a fun read. You'll know it when you see it.
That "really can't come about by accident"? I don't buy that.
I'll give you that it may well have been not by the "conventional" means. In fact, it's very hard to read useful data into a news report that merely describes the behavior of the bots, and doesn't say why they took the behavior, and especially a news report that uses phrases like "eerily wicked", "calmly", etc. to describe behaviors. But ok, let's give you at least that the evolutionary form had some effect.
Note well that in evolution, not every aspect of the machine evolves. Some parts are held constant. So an evolved model of something may have a learned behavior that is learned not because of the evolution but merely because of the situational structure and the basic learning mechanism. But, as I said, let's ignore that and assume you're right on this, because it will make my real point simpler.
Even if we knew evolution was used as the catalyzing mechanism, I think it would be an extraordinary claim if you were able to say that this is something that we were safe from seeing in a non-evolved robot that had the basic ability to evaluate and score its environment, select its goals, etc. I assume someone has done the (probably not difficult) legwork to show that this already implements a Turing machine. (I assume that mostly on the basis that a Turing machine isn't very complicated, and these robots sound moreso... I'd allege a good strong stake in the ground like "eerily wicked" implies Turing powerful. Though maybe someone is going to tell me that one of the evolutionary steps was provably to cross the Turing power line... I'm guessing it started already well across that line or it couldn't have performed at all, much less evolved, in the first iteration and/or it couldn't have reached this level by only 50 iterations. But it would depend even then on the initial construction and I'm willing to be surprised.)
But my point is that once Turing power is reached, and clearly we won't be stopping robots from doing that, I don't see how you can speak comfortably that this behavior won't happen due to regular neural nets. All you can say is that it hasn't. But I don't see anything fundamental about the computations involved in evolution that are likely to be any more elaborate than running a hypothetical, seeing the results, adjusting one's parameters, and trying again (vaguely like happens to the computer at the end of Wargames if you can stomach a somewhat whimsy visual metaphor in a serious discussion of computational complexity). People learn and re-learn processes, and I expect us to put that into robots. And even if we don't, the word "emergent" suggests that if it has survival value, it will end up happening.
It seems to me that it would be an extraordinary claim if you could really prove your statement
That would seem to me tantamount to a claim that there are two different kinds of intelligent processors and that you know a class of computations that can only be done on one and not on the other. To my knowledge, the ability to seriously prove such a claim is the holy grail of much investigation in computation science, and many hypothesize that it simply cannot be. But whether it can or cannot be, I don't think it's known now.
While this kind of stuff creeps me out as much as the next guy, and while it argues for being careful about what we trust robots to do, it's something we should know anyway because there many ways our trust can be violated without a robot lying. By far the more likely way they're going to let us down is just to exercise poor judgment. That is, to search for something that looks like a peanut butter sandwich but is really a rag with some grease on it... Getting the small details of common sense wrong is just as dangerous as anything deliberate.
What we really learn here is that the mathematics of learn things like lying as a strategy isn't remarkably complex; that is, (that is, the number of computational steps required to discover it works in at least some cases is small... note that we have no evidence that there is a general purpose intent to lie, only a case where communication was used and observed to score better in one mode than another). This is not a story about robots, it's a cautionary tale about neural nets, what they measure, how they fail, etc... and we didn't invent the idea of neural nets--we found it already installed in every living thing around us.
I went to the Museum of Science in Boston a few months back and saw, in the butterfly exhibit, a moth that had evolved coloration that was indistinguishable from an owl's face, hoping to scare off predators that were afraid of owls. Probably that's the more sophisticated result of the same notions. And yet it occurs in an animal that isn't, as a general purpose matter, a very sophisticated animal. Most people would find already-extant robots more socially engaging than a moth. For example, a moth is not capable of even serving up a beer during the game or vacuuming up the mess after your buddies go home.
So take heart: The likely truth is that this is unavoidable. If all it does is teach us to have a healthy skepticism for unrestrained technology, it's actually a good thing. We needed that skepticism anyway.
I agree with your remarks regarding emergent behavior and I liked the clock analogy, which I assume was in some ways a variant of Searle's Chinese Room.
I agree, too, that it's impossible to control. However, I think it's useful to protect oneself from it, just as it's useful to place limits on government.
Although it's also true that the term "it" is a little elusive here. Because, for example, it's easy to indulge the illusion that the "it" is the Internet, and that if you flick a switch, you're free of it. But the "it" is the "connected society", and it remains connected even if you turn off individual pieces of it. For example, whether you have the Internet or not, you can still be affected by identity theft, by insurance companies discrimintating based on genetic data, by someone in another country reading how to create an atomic bomb and then creating one, etc. Some of these were risks even before the Internet, when there were only paper libraries and paper telephone directories. The Internet has both accelerated that, and created combinatorial effects that are hard to quantify.
It's this full range of things I mean to attach when I say there should be protections. Not against data on a wire, but against the use of information. It can't be stopped, but it should be confronted as the real effect it is, and not ignored as if we must simply endure it.
Or from nowhere. The risk of a bad guy taking over is serious, but the risk that no one is at the helm is much more likely to lead us to death by Global Warming, for example.
You have to look no further than the US Congress to see a worked example. If you idealize every single member of Congress as intelligent, and I think a similar analogy can be made for people on the net or for companies on the net (where you still have to question intelligence sometimes, but let's not and say we did), it's pretty clear that the problem isn't just the sinister taking hold of someone with total power. It's also that it's easy to cause behavior that no one can take responsibility for, and that isn't in the best interest of individuals. The Internet is no different, but not because we didn't have examples of this before. Just because we didn't heed them.
Well, it would depend on whose definition of "free" you were using. This would satisfy your or my definition of free, but would have the problem that, absent source code, it wasn't GPL free. And if it came with source code, then putting back those lines and recompiling to get the full application wouldn't be hard. So anyone trying to protect their proprietary investment by just doing simple editing like this might not really be doing so.
It's not very relevant to this thread, but since you mention it: I am a strong advocate of the importance of copyright per se, for both programs and literature. Where I might have confused you is that I don't think copyrights on software published software that is 10-20 years old are really that valuable, and I think that society should change the duration of software copyright downward. The commercial shelf-life of software just isn't that long. I think a much greater public good is served by opening things up the possibility of reuse at that point. Books, by contrast, have a much longer commercial shelf-life--a well-written novel might sell for many years. So I can see a longer period justified based on the nature of the work that is coprighted. There are other objective differences between software and literature that I could cite as well, but this isn't the forum. Bottom line, though, I do believe and endorse the statement "copyright should exist".
The strongest argument for not reducing the duration of software copyright is that to the extent that old software might be resold without being first brought up to date, it risks re-introducing a lot of buggy and unsecure programs back into the world. In that regard, vendors may actually do a service by withholding such old and probably low-value programs out of some more personal motive.
As noted in another post about this article, it may be that there is no "format" other than "the code". If so, then the only free tool that is cheap to make is a wrapper around a complete application that just calls only part of that application. If so, making the wrapped tool free means giving away the entire program, not just the file part. In effect, then, this amounts to requesting that old versions be made free. Any difference between your proposal and asking old versions of their editor to just be made "free" (for whatever free you might be meaning) is just words, I suspect--nothing semantic.
Of course, this comes back to the question of whether there should be software patents at all, and whether software copyrights should have the immensely long durations that they do. Indeed, at some point, probably much shorter than happens now, having old tools be free so they can be recycled for other purposes may not be bad. It might even give vendors a kick in the pants to move faster to make newer tools be different enough that the old tools didn't threaten them. But bypassing a proper change in software copyright and patent law and instead just beating up on certain people who have things one wants does not seem the best approach to me.
If the application is useful, well-written, and well-documented, why would it matter who you went to? The problem with giving away source code is that this answer is far from obvious. What's most scary about your answer, in fact, is not the fact that you suggest this as a possible outcome--certainly it's possible--but that you strongly hint that there's no other possible answer.
Again, you are asserting this as if there is no alternative. And yet, again, with no pointer to hard data. As if somehow saying it were so will make it so.
To what degree of reliability? Based on what data? What if you're wrong? How would you rate your chances of being wrong?
Ah, ok. If it could access arbitrary pages, I was worried that it could access the admin component of my wireless router. (From the outside, WPA protects me from people accessing that admin stuff, but from the inside, it does not... only a more vanilla passwording protects me inside.) But if it's something like a limited access to the net, and if that limitation is not vulernable (nothing ever seems as safe as one expects), then it sounds like I'm fine.
I certainly do not have UPnP enabled now. My Belkin router fortunately comes with it off though offers the following remark in the online help, which given this recent news story I now see has a double-edged interpretation: “An application that is UPnP compliant has the ability to communicate with the Router, basically "telling" the Router which way it needs the firewall configured.”
I had just worried that some worm could find my gateway and enable it so that its friends could come in... but I read what you're saying as saying that if it got as far as being able to do that, it wouldn't need such an entry point anyway, it would already be inside and set up for business.
One follow-up question on this matter of only having access to the web site the item came from. A long time ago, I looked at security models in an informal way, but I am not an expert in this, so don't take my remarks about what I thought as any kind of authority--just as framing the question: I had thought that there was a subtle distinction between an applet on a page and a downloaded plug-in, and that it hinged on the question of whether the code was trusted. If Flash is a plug-in, not an applet, and is therefore trusted, and Flash is the thing that is doing the network connecting, then is Flash the thing that is holding my security in its hands, or does it have some way of making sure that the program it runs is subject to a more restrictive security model, such as I'm imagining applets to have? If I've asked the question in a way that makes it hard to answer because I've presupposed some division that's not there, feel free to repair the question before answering it.
Perhaps I should go read a recent reference on network security. But, sadly, there will be many with questions like these who don't have the time or training to do that, so hopefully the question lends itself to at least some sort of minimal high-level summary that doesn't require that.
Excuse my ignorance/confusion, but... I'm not up on the details of either Flash or UPnP, and yet I still need to understand this better and so I have a few questions.
Is the Flash being discussed the Flash player for a browser, right? (Not some sort of Flash related to flash memory and the BIOS and/or USB Flash drives? And the Flash issue is not in the router?)
Why is there a difference between the Flash vulnerability in different browsers? What's the basis of the protection? Is it because the player binaries differ between browsers, or because the security model of the browsers differ, or what?
If a Flash player is running malware already, why does it care any longer about the router? Isn't it already in my machine, and hence inside my network? And can't it generally get out quite easily with whatever data it finds without further problem? Or is there some security model limiting the actions of the Flash player to only certain operations?
Is it forbidden from writing files, particularly executable files? I assume a virus utility would notice this, but maybe since it's a trusted plug-in, it wouldn't?
If it can access web pages, isn't there also a potential vulnerability that many routers are configurable from inside the firewall over the network? In that case, couldn't it reenable UPnP itself? (Even if it was forbidden to read files from the disk and access the net, couldn't it just do the web page modification and then wait for a later copy of itself to arrive on a separate occasion to exploit the previously and silently opened hole?) If that can happen at all, will having a decent password for one's firewall reduce this risk? (Even though I have WPA-PSK enabled and a pretty long password, internal connections to a router over a secure connection seem like they're going to succeed because of the PSK, leaving the router's admin password the only thing in the way... or is there some other fortunate barrier?) Do routers tend to protect themselves from internal exhaustive or dictionary attacks? Would a virus protection tool notice this, or would it just think it normal that a browser was opening lots of web pages? In other words, do I need to switch my router to be configurable only over a serial link? (Even if I did, would I be vulnerable while the serial line was connected?)
If there's just a FAQ with answers to questions like these, please point me to it. I read the article, but it was pretty thick with device and protocol and program-specific jargon that even a technical person might not understand, depending on their areas of expertise.
I wouldn't bet my family's ability to eat on this. The problem isn't that there isn't money to be made on support. The problem is that a single person working his garage on software is not a support organization. Consequently, when it comes time to release that software, serious people (the kind with money) with a serious need for support (the kind they're going to bank their business on) won't trust a single individual for support. So your software will flow freely out, the support need may arise and support dollars may exist, but it's quite a gamble to assume they'll flow to you rather than to someone with a brand name and a committed resource of people that can stand behind a claim of support.
It's easy to want to believe you're going to get the money. But you are loads safer and much more likely to be right if you assume someone else will get it. You'd better assume you're going to get nothing and be satisfied with that. People may tell you otherwise, but I'd be surprised if they'd place money on that bet.
Advice, especially from evangelists, is cheap to offer when the one doing the offering doesn't have to deal with the consequences of being wrong. Don't let anyone convince you that doing something that is in the best interest of you and your own financial needs is some form of paranoia.
Of course, it's possible that the willingness of others to make free software has sufficiently driven down the price of software that there's no money you can make by selling it either. That's a different matter entirely. In that case, maybe you can sell service (supported by your own software, without releasing it), or maybe you just have something the world regards as worthless. That would be sad. But being sad would not make it impossible.
Part of the problem is: how does someone know it's free? Certainly the fact that it comes affixed with such a label doesn't tell you that--how do you know you're getting it from someone who is reliable if it comes in a free pack from anyone. If a van pulls up on the street and offers you something that looks like a consumer item at radically reduced price, you're already on alert... is the stuff stolen? You can get a lot of nearly free versions of DVDs of movies off of the net, too, but they turn out to be bootlegs. If people are charging media cost for disks of so-called free software, software that obviously took time to produce, how is that not going to look the same to someone who isn't technically savvy? Sure, to someone who is comfortble browsing Sourceforge it might work, but to that person's mom?
The trust issue is a serious one, and the open source community deals with that by a set of mechanisms that involve finding trusted sources. But just handing someone a disk and saying "you can use this" is not it. It requires technical and legal sophistication to know you're not becoming (or just paying) a criminal by buying. It also requires fair technical sophistication to know where to get updates, etc...
You're not necessarily doing a non-technical person a favor by getting them into this. It's like getting someone to buy a car that has to be maintained oneself instead of getting them to buy a name brand with people that can maintain it. That's great for people who like being tinkerers, but not for everyone. And I'd think most people who can't fix their own car would be pretty suspicious of buying from folks who give away cars for free and said "don't worry about the cost, I'll make my money on the maintenance."
Are we talking per-customer watermarks? (The article didn't seem to say.) Aside from the usual privacy implications, that would have its own problems, since it would allow for unbounded downstream prosecution of anyone who ever let even one copy go free, including through malware. It would make it quite a liability to even buy such stuff.
I have to wonder why everyone's trying to put a band-aid on the problem.
Because politics rarely involves one person getting their way, no matter how good. So the practical person, for better or worse, examines compromises.
I'm actually seriously with you on that. Lots of people fear nuclear. But if you really believe either he global warming issue, or the fact that we're peaking on oil, we're going to need cheap, clean energy soon. I'm all for wind and solar, but don't expect that to deploy in enough time to really work. So your band-aid (heh--see, everyone has one--it's only natural) of filling the gap with a bunch of nuclear plants sounds like the only way forward that makes much sense to address the cause, not the symptom... at least in the next 10-20 years.
Just, please, let's put them above the plain that might be flooded by global warming. And definitely not in the basement of anyone with a government-controlled thermostat.