Maybe not. This (old) article explains that the ISS can continue to be supported with a reduced shuttle fleet. Bizarrely, it even assumes that there WILL be another shuttle disaster, just like today.
Manned space flight (both shuttle trips, and the International Space Station) are today worth neither the risk nor the money. I like what John Pike said about the ISS: "The value of the science that can be done on the Space Station is trivial compared to the cost of the Space Station. Piloted spaceflight is about politics."
Let's look specifically at the ISS, which is the destination for most of the recent shuttle flights. Keeping humans supplied in space takes many extra trips up and down: all the air, water, food, living space, and exercise equipement takes up valuable cubic meters. And all of the provisions for safety and gentle re-entry further reduce the fuel efficiency of the rockets.
The ISS program, and the supply flights to build & support it, will have a total price tag of at around $100,000,000,000.
Scientific-notation kinds of fundage ($1e11)!! You'd have to be a NASA researcher just to count it all.
Virtually all of the science and maintenannce done on Shuttles and the ISS could be accomplished by semi-autonomous robots. Sure, today maybe our robotics and AI technology isn't good enough to substitute for some of the tricky things where a dynamic, flexible human is needed. Well, try investing a fraction of the $1e11 budget into researching those systems, and then tell me how well they work!
Developing better robots to operate space equipment won't only make extra-planetary research safer and cheaper- it'll also produce technological advances that will benefit civilians around the world!
(Rocket-boosters are only needed by astronauts and admirals. But reliable robot manipulators could be useful to anyone)
I fear for the public reaction agenst NASA and space traval from this day forward.
I hope the public wises up that manned space flight is an expensive and dangerous form of esteem-boosting entertainment.
Re:Basic AI research important
on
Infinite Games?
·
· Score: 1
My whole point above is that academic fields experience scope-creep, as you've abundantly illustrated. AI is a subset of CS- just because an AI professor writes about something doesn't mean it's in his specific subfield.
in some cases a trivial task compared to the over-arching goal of developing some ad-hoc reasoning within an agent.
That's possibly the broadest definition of "trivial" I've ever encountered.
Ok, here's a fun game: define "agent". "AI" is straightforward in comparison. Sure, there is a traditional set of problem-solving approaches associated with "agent systems", but the definitions of agent I've seen are either plainly wrong, or expansive enough to cover nearly any piece of software.
but genetic algorithms are an AI concept
Genetic algorithims are a CS concept! (As is A* search, and other things you've listed) When GA were invented (Bremermann 1962) they were connected to theories of computation and information, not artificial intelligence. The mere fact that schools like to lump it into the AI curricula doesn't rewrite history.
Quite often, authors will lump AI and GA (and other things) into a field called "Artificial Life", as a way of acknowledging that GA and GP aren't part of AI, but may be interesting to the same audience. (Koza '92, for instance, is like this)
Would you call exhaustive or random searches of a problem space AI? Of course not, they're fundmental CS techniques (employed, of course, as a jumping off point to learn how to avoid getting stuck in perebor). But the output of a GA run of M generations of N individuals is strictly equivalent to the output of a random search of a much larger number of samples (about N! * M * survival_fraction, or something like that).
It's a dumb search. It doesn't deserve the name "intelligent". It may be effective, but it's still dumb.
Planting mines beforehand truely makes it easy. A shotgun blast to the back of the neck will do the job as well- except the detonation when her gear and powerpack blows up will take your legs off!
(If you don't have healthkits, you can still crawl all across the airport to your egress, but it takes 45 minutes)
Re:Translation: AI is Nowheresville
on
Infinite Games?
·
· Score: 1
You said nothing about my central point- that by defining Darwinian processes of natural selection as intelligent, you have then claimed that the entire planet earth (and indeed possibly the universe) are intelligent beings.
The Gaia hypothesis is amusing, but it's got a flaky reputation.
And since it forms an entire course at Edinburgh University Dept. of AI where I study
"AI" departments, having noticed that AI won't happen for decades yet, have grabbed on whatever related subjects they can to get some results in the meantime.
In the case of GA, artificial intelligence is not used- the real kind is. To get a good solution to a problem with GA, you need to apply a significant amount of human intelligence in creating a clever, feasible fitness function. Intuition about the problem domain really helps. Get it a little bit wrong, and your GA will either settle on a local minima, or converge tractably slowly.
Hive intelligence for example... no single ant in a colony has any clue as to what is going on, and there's no controlling insect, yet the hive has intelligence when viewed as a whole.
Specious.
You can say that about anything. "No single neuron/relay in a brain/microprocessor has any clue what's going on..."
And on the other hand, I've got some pet honeybees (they're hibernating right now), but let me assure you, individual bees are quite intelligent. They can talk, read, communicate geometric directions. That should be enough to score IQs of 10+ points.
(A full spectrum IQ test can only be performed by psychologists, and can cover the full range of intelligence from brain-damage to supergenuius. Twitching upon being touched is worth a point)
Re:Translation: AI is Nowheresville
on
Infinite Games?
·
· Score: 1
Wait a second. You're telling me that a genetic algorithim (evolutionary evaluation) is part of Artificial Intelligence?
Does this mean that the isn't really a conflict between Darwinists and "Intelligent Design" proponents?
(Genetic Algorithims are all about solving complex problems without any form of intelligence. They're just a performance optimization on a random search.)
Re:Didn't NOLF2 implement something like this alre
on
Infinite Games?
·
· Score: 1
Nope, nothing like this.
The behaviors you mention are AI to simulate characters in the game. The behaviors pretend like they're guards for international criminals, and react appropriately. Thought processing goes "I want to defend my hideout. I heard noise from downstairs. Nobody's authorized in here! I'll start shooting"
The topic of this article is using AI to simulate a character outside the game. A virtual storyteller. Theoretically, if this worked well, the thought processing would be along the lines of "I want to demonstrate that princples can be expensive. So I'll pick an enemy the player has resisted, and a character sympathetic to the player, and have the enemy hurt/threaten the friend. Now, which NPCs that the player has interacted with best fit those roles?"
From a Dungeons and Dragons perspective, the first approach is for an AI to play "Silvidyn, the Dark Elf Warlord" and the second is for it to play "Willie Jansen, the Dungeon Master".
Re:Slashdot is aiming way too low
on
Infinite Games?
·
· Score: 1
but almost none know where to find presentations about story-AI at an introductory level.
You're too modest. Robotwisdom.com is such a seminal webpage for the blogging pheonmenon that everyone must've visited it by now! (and then they've all left, because everyone prefers GIF pictures of text to the HTML text itself)
You say that as if the BBC article offered one, but at best it offered a link to some PDFs that claim to be one.
I'll give you that aside from links, the BBC provides about 75 words of interesting content. But since AI-guided games haven't been released yet, text of research papers is the best anybody can offer. A new bundle of papers is provided by the Mimesis / Liquid Narrative crew, and they're not linked from your website yet, so we're getting some new service from this article.
If you feel the NCSC publications are wrong or redundant, it would be interesting to see some examples of why. Lacking such specific connections to the article, there's no real point to proposing your page as alternative reading.
(It's not a bad page, but Aug99 is getting stale in terms of staying abreast of alleged recent developments)
Re:Basic AI research important
on
Infinite Games?
·
· Score: 1
Actually, you're wrong
Actually, you've just redefined AI to be an easier problem. AI can't work yet, so you've picked some related research and "promoted" it to be called AI.
"Intelligence" is the ability to succeed on an Intelligence Test, like an IQ test or SAT. Until AI starts to get reasonable scores on those (or the more field-specific Turing Test), it's not AI.
AI is about allowing computers to use concepts approximating human reasoning
"Articifical" "Intelligence". Anything else is incorrect. The expansive definition may be more useful for you in your research, but it is linguistically wrong. Your examples only use parts of a potential future AI system.
This pheomenon is fairly common, and a little interesting. A new thing is invented, with a new name, and a professional class springs up to study and improve it. As they work, they decide to focus on something other than what the word originally meant, so they create a new definition for their own use.
Look at "tank" (a military vehicle). Originally (and to the public, still) it meant "a heavily armored ground vehicle". (As developed in 1918 to survive machinegun fire) But today's professional soldiers define it as "a heavily armored ground vehicle, with weapons capable of destroying something similarly armored". Thus they say that the M1A1 is a tank, but not the M2A3
AI isn't just about making things act like humans. That's only good for human-computer interaction (...) and entertainment
No, there are many more reasons to make a computer act like a human. Economic predictions, military simulations... I won't insult you by listing more.
then you have to fight gunther/hermann/whatever-german-name, but you can't win because he is invincible
The worst part about that section is what can happen to your equipment. Throughout the whole game, your ammunition is painfully limited. ("I'm a top secret government assasin! Can't you guys give me a full clip of 9mm ammo before each mission??")
If you honestly try to fight past Gunther and his armored police squad, you can waste all your valuable ammunition on him, (and eat all your healthpacks) making the next part of the game much harder. (You break out of prison, where they had conveniently let you keep your bullets)
The guy you were ordered to kill was a rebel financier (some Russian, you find him on an airplane).
There IS a fork there, albeit a minor one. You can kill him, or if you don't an NPC tries to. You can let this happen, or you can stop her. (Which results in you two fighting to the death). After she's dead, you explain to UNATCO that she was shot by terrorists, and you continue in your job there.
(If you didn't kill her, then she attacks you 2 missions later regardless).
Because it all happens so fast, most players don't realize they had any choice in the matter. Also, there's a 3rd option the game didn't plan for: you have a full range of non-lethal weaponry, and can easily capture the guy and drag him back to HQ. The NPCs have no awareness that you've done this, though, and will assume he's dead.
OTOH, there is a scene later on, where you're ordered to "arrest" your brother. But you absolutely cannot obey that order- he's invulnerable to your attacks. You can choose to help him escape (difficult, but possible), or you can allow the other cops to kill him. But you simply cannot take him down yourself. This is the single biggest potential fork-point in the game, and they missed it.
Re:Slashdot is aiming way too low
on
Infinite Games?
·
· Score: 1
It's polite to wrap posts like that with a shameless plug tag. Also it helps to at least paste in a few lines from your site which are particularly relevant. Otherwise, your few visitors will immediately flee when they see a bunch of links to LISP, Mania, and Cyc, with no apparent attachment to video games.
and an insult to Slashdot's readership
You must read at +5 or something. That's a drastic overestimation.
infinitely more perspective
We're not looking for "perspective", background, or topics of academic interest. People are intersested in results. (Those few who enjoy abstract theory already know where to find it, and need neither the BBC nor Slashdot to remind them)
Plenty has been written about these topics, but AI-directed narratives have never had a successful practical application yet. You might say that it's only a matter of computational power or incremental work before we achieve a "CPU gamemaster", but until then, every concrete step along the way will be interesting.
Empirically, though, Tetris (and related games in the abstract/geometric puzzle genre) have been shown to be more "interesting". The total quantity of human interest they have accumulated, measured in hours of play, is hundreds of times greater than FPS and Survival Horror have achieved.
When the gameplay is just delaying factors on the way to viewing a precreated story, it's an exclusionary barrier- only people who like the story AND have the skills to reach it will play. Many of them won't get all the way through. The storyline is finite, though. At some point it will end, and then only a rare breed of aficionados bothers to play through again.
Puzzle games don't have their playability constrained by the quality or duration of any prewritten story. It's all about the game.
What is the definition of game? In many fields (from computer science, to sports) it means that players are taking actions and making choices to influence an outcome. It's completely different from a story. And it certainly doesn't mean "complete little dexterity challenges so we can be sure you're still awake before we show the next part of the story".
You've argued that Half-Life and Resident Evil are better stories than Tetris. That is a given, and it is irrelevant to the quality of the game.
Take a game like Half-Life
Ok. The story of Half-Life is "Secret scientists for the evil government foul up a teleportation experiment and release hordes of zombies and alien monsters into their hidden base. A lone survivor takes up arms to defeat them."
Wow! The exact same story as used in Doom. And Quake too (if anyone noticed). Sure, Half-Life does a better job of presenting the same old events, but few serious authors ("story professionals") would dignify it would the term "story".
The major twist Half-Life added was the presents of the Army forces- but that actually weakened the storytelling significantly. Those intractable enemy forces were completely unbelievable. "Hello, I'm a 19-year old private who's just been attacked by carnivorous horrors from Dimension-X. I think I'll scrawl some grafitti with the name of a random local lab assistant I haven't even heard of!"
It makes sense that you can't parley with a demon on alien freak, but being unable to even try talking sense into the soldiers emphasizes the limitations of the game interaction. If it had been a movie, the hero wouldn't have put a bullet into the head of every soldier he saw. He'd have evaded them for a while until the aliens overwhelmed the unprepared soldiers, then rescued the survivors and recruited them as backup for the final battle.
The funniest part is when a soldier is attacked by a "barnacle". The slimey tentacle wraps around his neck, choking him as he's dragged inexorably into the crushing teeth... But a mysterious stranger in exotic armor bursts from the shadows and roasts the monster with a blast of controlled fusion energy. The rescued marine tumbles to the ground- And before even hitting the ground, has opened fire on his savior!
You're suggestion to 'just remove the worm' would give the author notice that the feds were on to him.
Or that Norton/MacAfee/Microsoft was on to him. Or he might think the sysadmin was on to him. Or that the user had randomly reinstalled windows. Or he'd even forget he'd ever hit that computer.
That's like just painting over grafitti everyday instead of preventing it
Invalid comparison.
When performing grafitti, the perp need physical proximity to the target. Therefore physical protection (a cop on patrol) can be effective.
To write a worm, you needn't be anywhere near the target. Therefore protections which boil down to "pull out your gun and grab him" will not be very effective.
This article showed us that even in the UK (the US's biggest lackey-state), the FBI can't get the prosecutions it wants. We shouldn't expect arrest to be a much more effective deterent through the rest of the (US-antagonist) world.
Jokers say that Linux contributors are doing free development for IBM. So now the US Government is doing free research for Microsoft.
The question is will the "Feds" be at least somewhat successful in their attempts to thwart future worms and other virii?
The answer is no. By squelching this "attack" (if they really did), they've just allowed Joe Public to continue postponing learning about putting his money into secure computer systems.
Following recent testimony, it has come to our attention that Microsoft(tm) products perform mission-critical operations in our national War Against Terror(tm).
Consequently, the source code for Microsoft Windows(tm), Microsoft Office(tm), Microsoft Bob(tm), and related software, is immediately upgraded to a top-secret classification.
Federal Marshals will be arriving shortly to quaranty your facility, until the NSA can complete background checks on each of your personnel to ensure they can be trusted with such a grave responsibility.
Non-citizens, or those failing background checks, will be interred as an enemy combatant until the cessation of the conflict.
Apparently he observed that the current approach to fighting "cybercrime" is (instead of building safe networks) to sic the FBI on them, and that this won't work if the attacker is outside of our jurisdiction.
Therefore he hopes that really soon now, our jurisdiction can expand to swallow the entire planet.
It's hard, but it must be done. Fine, it can take 5-7 years, but it needs to happen. (Swapping out some software is trivial in comparison to things like airport security and National Missile Defense)
The problem these stories show us is that the Federal Cybercops are spending all their effort to barely, occasionally control unfocused, amateur miscreants. Pranksters out for fun. "cybercrime"
They should be hardening against attacks by state-sponsored saboteurs who are trained, funded, organized and motivated. Enemies who won't submit to arrest, and who won't flinch at B&E of a Colonel's house to bug his laptop. (Or take his password at gunpoint.) The attack won't be tentative or experimental- it won't come until the assailants are ready to apply it in force. "cyberwar"
The government can't even keep casual "cybercrime" in check, inspiring no confidence that they'll do much better in a "cyberwar", which should be their main concern. (They've recently used the word "cyberterrorism", which only confuses matters)
Their current approach just creates a false sense of security. The sooner they scale it back, the sooner the public will start to demand & install truely secure computing, and the safer we'll all be.
Private parties can, at most, disperse the botnet or terminate the attacker's account.
Some of those private parties are software developers, who can do a little more- they can fix insecurities, and prevent them from happening in the first place. The only longterm solutions to vulnerabilities.
So far, though, it seems that developers (meaning primarily Microsoft) still don't pay enough attention to security.
Why not? Because the marketplace doesn't value secure software, so they aren't punished for not providing it.
Why doesn't the market value security? Because they think government departments like the one described will protect them, instead of relying on their software vendor or themselves.
By providing these hardworking "cybercrime" specialists, the government accomplishes 3 things:
Reduce the economic infrastructure's future resistant to future attacks based in foreign countries. The FBI has little jurisdiction in South Korea, and none to speak of in China.
I'm not saying that no crime committed on a computer should be punished- but that both the level of effort put into hunting, and the amount of punishment allocated should be reduced.
Re:Linux games vs. shareware stuff for Win
on
25 Best Linux Games
·
· Score: 1
You didn't figure it out?
Type "xyzzy" and then a single pixel in the corner of your screen will blink when you mouseover a mine.
Most all school projects are "extra work", since students can rarely forge new ground that hasn't been investigated before. Instead, they learn to correctly duplicate known experiments, to get up to SOTA before following their own ideas.
In this case, the machine-vision problem of recognizing blocks with a camera (which inserts distortion and uncertainty to the image) is an interesting challenge on its own.
While viewing a computer screen is easier than most MV problems (the target is self-illuminated, so variable lightsource angles are less of a problem), it could have useful future applications. Especially if someday we're beset by hardened DRM computers that don't allow you to interact programatically- mouse input and LCD output only! Then the practice of watching a screen with the camera of your PDA might become the only way to take controlled files off those machines.
(Watch the "Ghost in the Shell" movie for bizarre examples of this)
Slashdot Mirror
Maybe not. This (old) article explains that the ISS can continue to be supported with a reduced shuttle fleet. Bizarrely, it even assumes that there WILL be another shuttle disaster, just like today.
http://www.saber.net/~donaldrf/buildss.html
(Search for "WHEN A SHUTTLE IS LOST")
"Is space worth the risk?"
No, it's not. More precisely, manned space travel isn't worth the risk. (Unmanned missions are risk-free by comparison)
Just look at the kinds of leading edge science this crew died to perform:
http://www.wff.nasa.gov/~sspp/sem/about.html
Manned space flight (both shuttle trips, and the International Space Station) are today worth neither the risk nor the money. I like what John Pike said about the ISS: "The value of the science that can be done on the Space Station is trivial compared to the cost of the Space Station. Piloted spaceflight is about politics."
Let's look specifically at the ISS, which is the destination for most of the recent shuttle flights. Keeping humans supplied in space takes many extra trips up and down: all the air, water, food, living space, and exercise equipement takes up valuable cubic meters. And all of the provisions for safety and gentle re-entry further reduce the fuel efficiency of the rockets.
The ISS program, and the supply flights to build & support it, will have a total price tag of at around $100,000,000,000.
Scientific-notation kinds of fundage ($1e11)!! You'd have to be a NASA researcher just to count it all.
Virtually all of the science and maintenannce done on Shuttles and the ISS could be accomplished by semi-autonomous robots. Sure, today maybe our robotics and AI technology isn't good enough to substitute for some of the tricky things where a dynamic, flexible human is needed. Well, try investing a fraction of the $1e11 budget into researching those systems, and then tell me how well they work!
Developing better robots to operate space equipment won't only make extra-planetary research safer and cheaper- it'll also produce technological advances that will benefit civilians around the world!
(Rocket-boosters are only needed by astronauts and admirals. But reliable robot manipulators could be useful to anyone)
I fear for the public reaction agenst NASA and space traval from this day forward.
I hope the public wises up that manned space flight is an expensive and dangerous form of esteem-boosting entertainment.
My whole point above is that academic fields experience scope-creep, as you've abundantly illustrated. AI is a subset of CS- just because an AI professor writes about something doesn't mean it's in his specific subfield.
in some cases a trivial task compared to the over-arching goal of developing some ad-hoc reasoning within an agent.
That's possibly the broadest definition of "trivial" I've ever encountered.
Ok, here's a fun game: define "agent". "AI" is straightforward in comparison. Sure, there is a traditional set of problem-solving approaches associated with "agent systems", but the definitions of agent I've seen are either plainly wrong, or expansive enough to cover nearly any piece of software.
but genetic algorithms are an AI concept
Genetic algorithims are a CS concept! (As is A* search, and other things you've listed)
When GA were invented (Bremermann 1962) they were connected to theories of computation and information, not artificial intelligence. The mere fact that schools like to lump it into the AI curricula doesn't rewrite history.
Quite often, authors will lump AI and GA (and other things) into a field called "Artificial Life", as a way of acknowledging that GA and GP aren't part of AI, but may be interesting to the same audience. (Koza '92, for instance, is like this)
Would you call exhaustive or random searches of a problem space AI? Of course not, they're fundmental CS techniques (employed, of course, as a jumping off point to learn how to avoid getting stuck in perebor). But the output of a GA run of M generations of N individuals is strictly equivalent to the output of a random search of a much larger number of samples (about N! * M * survival_fraction, or something like that).
It's a dumb search. It doesn't deserve the name "intelligent". It may be effective, but it's still dumb.
Planting mines beforehand truely makes it easy.
A shotgun blast to the back of the neck will do the job as well- except the detonation when her gear and powerpack blows up will take your legs off!
(If you don't have healthkits, you can still crawl all across the airport to your egress, but it takes 45 minutes)
You said nothing about my central point- that by defining Darwinian processes of natural selection as intelligent, you have then claimed that the entire planet earth (and indeed possibly the universe) are intelligent beings.
..."
The Gaia hypothesis is amusing, but it's got a flaky reputation.
And since it forms an entire course at Edinburgh University Dept. of AI where I study
"AI" departments, having noticed that AI won't happen for decades yet, have grabbed on whatever related subjects they can to get some results in the meantime.
In the case of GA, artificial intelligence is not used- the real kind is. To get a good solution to a problem with GA, you need to apply a significant amount of human intelligence in creating a clever, feasible fitness function. Intuition about the problem domain really helps. Get it a little bit wrong, and your GA will either settle on a local minima, or converge tractably slowly.
Hive intelligence for example... no single ant in a colony has any clue as to what is going on, and there's no controlling insect, yet the hive has intelligence when viewed as a whole.
Specious.
You can say that about anything. "No single neuron/relay in a brain/microprocessor has any clue what's going on
And on the other hand, I've got some pet honeybees (they're hibernating right now), but let me assure you, individual bees are quite intelligent. They can talk, read, communicate geometric directions. That should be enough to score IQs of 10+ points.
(A full spectrum IQ test can only be performed by psychologists, and can cover the full range of intelligence from brain-damage to supergenuius. Twitching upon being touched is worth a point)
Wait a second. You're telling me that a genetic algorithim (evolutionary evaluation) is part of Artificial Intelligence?
Does this mean that the isn't really a conflict between Darwinists and "Intelligent Design" proponents?
(Genetic Algorithims are all about solving complex problems without any form of intelligence. They're just a performance optimization on a random search.)
Nope, nothing like this.
The behaviors you mention are AI to simulate characters in the game. The behaviors pretend like they're guards for international criminals, and react appropriately. Thought processing goes "I want to defend my hideout. I heard noise from downstairs. Nobody's authorized in here! I'll start shooting"
The topic of this article is using AI to simulate a character outside the game. A virtual storyteller. Theoretically, if this worked well, the thought processing would be along the lines of "I want to demonstrate that princples can be expensive. So I'll pick an enemy the player has resisted, and a character sympathetic to the player, and have the enemy hurt/threaten the friend. Now, which NPCs that the player has interacted with best fit those roles?"
From a Dungeons and Dragons perspective, the first approach is for an AI to play "Silvidyn, the Dark Elf Warlord" and the second is for it to play "Willie Jansen, the Dungeon Master".
but almost none know where to find presentations about story-AI at an introductory level.
You're too modest. Robotwisdom.com is such a seminal webpage for the blogging pheonmenon that everyone must've visited it by now! (and then they've all left, because everyone prefers GIF pictures of text to the HTML text itself)
You say that as if the BBC article offered one, but at best it offered a link to some PDFs that claim to be one.
I'll give you that aside from links, the BBC provides about 75 words of interesting content. But since AI-guided games haven't been released yet, text of research papers is the best anybody can offer. A new bundle of papers is provided by the Mimesis / Liquid Narrative crew, and they're not linked from your website yet, so we're getting some new service from this article.
If you feel the NCSC publications are wrong or redundant, it would be interesting to see some examples of why. Lacking such specific connections to the article, there's no real point to proposing your page as alternative reading.
(It's not a bad page, but Aug99 is getting stale in terms of staying abreast of alleged recent developments)
Actually, you're wrong
Actually, you've just redefined AI to be an easier problem. AI can't work yet, so you've picked some related research and "promoted" it to be called AI.
"Intelligence" is the ability to succeed on an Intelligence Test, like an IQ test or SAT. Until AI starts to get reasonable scores on those (or the more field-specific Turing Test), it's not AI.
AI is about allowing computers to use concepts approximating human reasoning
"Articifical" "Intelligence". Anything else is incorrect. The expansive definition may be more useful for you in your research, but it is linguistically wrong. Your examples only use parts of a potential future AI system.
This pheomenon is fairly common, and a little interesting. A new thing is invented, with a new name, and a professional class springs up to study and improve it. As they work, they decide to focus on something other than what the word originally meant, so they create a new definition for their own use.
Look at "tank" (a military vehicle). Originally (and to the public, still) it meant "a heavily armored ground vehicle". (As developed in 1918 to survive machinegun fire) But today's professional soldiers define it as "a heavily armored ground vehicle, with weapons capable of destroying something similarly armored". Thus they say that the M1A1 is a tank, but not the M2A3
AI isn't just about making things act like humans. That's only good for human-computer interaction (...) and entertainment
No, there are many more reasons to make a computer act like a human. Economic predictions, military simulations... I won't insult you by listing more.
Well, obviously he's Boromir the 2nd, first son of Faramir, and named for his brave, departed uncle.
He must've taken the throne after Aragorn failed to sire an heir.
then you have to fight gunther/hermann/whatever-german-name, but you can't win because he is invincible
The worst part about that section is what can happen to your equipment. Throughout the whole game, your ammunition is painfully limited. ("I'm a top secret government assasin! Can't you guys give me a full clip of 9mm ammo before each mission??")
If you honestly try to fight past Gunther and his armored police squad, you can waste all your valuable ammunition on him, (and eat all your healthpacks) making the next part of the game much harder. (You break out of prison, where they had conveniently let you keep your bullets)
spoilers coming!! ... ... ...
The guy you were ordered to kill was a rebel financier (some Russian, you find him on an airplane).
There IS a fork there, albeit a minor one. You can kill him, or if you don't an NPC tries to. You can let this happen, or you can stop her. (Which results in you two fighting to the death). After she's dead, you explain to UNATCO that she was shot by terrorists, and you continue in your job there.
(If you didn't kill her, then she attacks you 2 missions later regardless).
Because it all happens so fast, most players don't realize they had any choice in the matter. Also, there's a 3rd option the game didn't plan for: you have a full range of non-lethal weaponry, and can easily capture the guy and drag him back to HQ. The NPCs have no awareness that you've done this, though, and will assume he's dead.
OTOH, there is a scene later on, where you're ordered to "arrest" your brother. But you absolutely cannot obey that order- he's invulnerable to your attacks. You can choose to help him escape (difficult, but possible), or you can allow the other cops to kill him. But you simply cannot take him down yourself. This is the single biggest potential fork-point in the game, and they missed it.
It's polite to wrap posts like that with a shameless plug tag. Also it helps to at least paste in a few lines from your site which are particularly relevant. Otherwise, your few visitors will immediately flee when they see a bunch of links to LISP, Mania, and Cyc, with no apparent attachment to video games.
and an insult to Slashdot's readership
You must read at +5 or something. That's a drastic overestimation.
infinitely more perspective
We're not looking for "perspective", background, or topics of academic interest. People are intersested in results. (Those few who enjoy abstract theory already know where to find it, and need neither the BBC nor Slashdot to remind them)
Plenty has been written about these topics, but AI-directed narratives have never had a successful practical application yet. You might say that it's only a matter of computational power or incremental work before we achieve a "CPU gamemaster", but until then, every concrete step along the way will be interesting.
Tetris as the most interesting game? Hardly!!
Empirically, though, Tetris (and related games in the abstract/geometric puzzle genre) have been shown to be more "interesting". The total quantity of human interest they have accumulated, measured in hours of play, is hundreds of times greater than FPS and Survival Horror have achieved.
When the gameplay is just delaying factors on the way to viewing a precreated story, it's an exclusionary barrier- only people who like the story AND have the skills to reach it will play. Many of them won't get all the way through. The storyline is finite, though. At some point it will end, and then only a rare breed of aficionados bothers to play through again.
Puzzle games don't have their playability constrained by the quality or duration of any prewritten story. It's all about the game.
What is the definition of game? In many fields (from computer science, to sports) it means that players are taking actions and making choices to influence an outcome. It's completely different from a story. And it certainly doesn't mean "complete little dexterity challenges so we can be sure you're still awake before we show the next part of the story".
You've argued that Half-Life and Resident Evil are better stories than Tetris. That is a given, and it is irrelevant to the quality of the game.
Take a game like Half-Life
Ok. The story of Half-Life is "Secret scientists for the evil government foul up a teleportation experiment and release hordes of zombies and alien monsters into their hidden base. A lone survivor takes up arms to defeat them."
Wow! The exact same story as used in Doom. And Quake too (if anyone noticed). Sure, Half-Life does a better job of presenting the same old events, but few serious authors ("story professionals") would dignify it would the term "story".
The major twist Half-Life added was the presents of the Army forces- but that actually weakened the storytelling significantly. Those intractable enemy forces were completely unbelievable. "Hello, I'm a 19-year old private who's just been attacked by carnivorous horrors from Dimension-X. I think I'll scrawl some grafitti with the name of a random local lab assistant I haven't even heard of!"
It makes sense that you can't parley with a demon on alien freak, but being unable to even try talking sense into the soldiers emphasizes the limitations of the game interaction. If it had been a movie, the hero wouldn't have put a bullet into the head of every soldier he saw. He'd have evaded them for a while until the aliens overwhelmed the unprepared soldiers, then rescued the survivors and recruited them as backup for the final battle.
The funniest part is when a soldier is attacked by a "barnacle". The slimey tentacle wraps around his neck, choking him as he's dragged inexorably into the crushing teeth...
But a mysterious stranger in exotic armor bursts from the shadows and roasts the monster with a blast of controlled fusion energy. The rescued marine tumbles to the ground-
And before even hitting the ground, has opened fire on his savior!
I would think that the penultimate scripted virus should be one that, like the genetic variety,
Why would that be the next-to-last virus ever created?
A script is basically variable names and values.. well what if you could randomly generate variable names, of
Actually at some point, the script will really need to come down to system calls. Just assigning variables all day long won't accomplish much.
search-and-replace to generate copy Y with the new names.
Both this trick, and countermeasures like you mentioned, have been happening for years. Or does the phrase "polymorphic virus" not ring a bell?
You're suggestion to 'just remove the worm' would give the author notice that the feds were on to him.
Or that Norton/MacAfee/Microsoft was on to him. Or he might think the sysadmin was on to him. Or that the user had randomly reinstalled windows. Or he'd even forget he'd ever hit that computer.
That's like just painting over grafitti everyday instead of preventing it
Invalid comparison.
When performing grafitti, the perp need physical proximity to the target. Therefore physical protection (a cop on patrol) can be effective.
To write a worm, you needn't be anywhere near the target. Therefore protections which boil down to "pull out your gun and grab him" will not be very effective.
This article showed us that even in the UK (the US's biggest lackey-state), the FBI can't get the prosecutions it wants. We shouldn't expect arrest to be a much more effective deterent through the rest of the (US-antagonist) world.
Oh yes, very beneficial.
Jokers say that Linux contributors are doing free development for IBM. So now the US Government is doing free research for Microsoft.
The question is will the "Feds" be at least somewhat successful in their attempts to thwart future worms and other virii?
The answer is no. By squelching this "attack" (if they really did), they've just allowed Joe Public to continue postponing learning about putting his money into secure computer systems.
Attention Mr. Gates:
Following recent testimony, it has come to our attention that Microsoft(tm) products perform mission-critical operations in our national War Against Terror(tm).
Consequently, the source code for Microsoft Windows(tm), Microsoft Office(tm), Microsoft Bob(tm), and related software, is immediately upgraded to a top-secret classification.
Federal Marshals will be arriving shortly to quaranty your facility, until the NSA can complete background checks on each of your personnel to ensure they can be trusted with such a grave responsibility.
Non-citizens, or those failing background checks, will be interred as an enemy combatant until the cessation of the conflict.
Sincerely,
F. B. I.
It doesn't sound like he was kidding.
Apparently he observed that the current approach to fighting "cybercrime" is (instead of building safe networks) to sic the FBI on them, and that this won't work if the attacker is outside of our jurisdiction.
Therefore he hopes that really soon now, our jurisdiction can expand to swallow the entire planet.
Pax Americana at last!
It's hard, but it must be done. Fine, it can take 5-7 years, but it needs to happen. (Swapping out some software is trivial in comparison to things like airport security and National Missile Defense)
The problem these stories show us is that the Federal Cybercops are spending all their effort to barely, occasionally control unfocused, amateur miscreants. Pranksters out for fun.
"cybercrime"
They should be hardening against attacks by state-sponsored saboteurs who are trained, funded, organized and motivated. Enemies who won't submit to arrest, and who won't flinch at B&E of a Colonel's house to bug his laptop. (Or take his password at gunpoint.) The attack won't be tentative or experimental- it won't come until the assailants are ready to apply it in force.
"cyberwar"
The government can't even keep casual "cybercrime" in check, inspiring no confidence that they'll do much better in a "cyberwar", which should be their main concern. (They've recently used the word "cyberterrorism", which only confuses matters)
Their current approach just creates a false sense of security. The sooner they scale it back, the sooner the public will start to demand & install truely secure computing, and the safer we'll all be.
Some of those private parties are software developers, who can do a little more- they can fix insecurities, and prevent them from happening in the first place. The only longterm solutions to vulnerabilities.
So far, though, it seems that developers (meaning primarily Microsoft) still don't pay enough attention to security.
Why not? Because the marketplace doesn't value secure software, so they aren't punished for not providing it.
Why doesn't the market value security? Because they think government departments like the one described will protect them, instead of relying on their software vendor or themselves.
By providing these hardworking "cybercrime" specialists, the government accomplishes 3 things:
I'm not saying that no crime committed on a computer should be punished- but that both the level of effort put into hunting, and the amount of punishment allocated should be reduced.
You didn't figure it out?
Type "xyzzy" and then a single pixel in the corner of your screen will blink when you mouseover a mine.
Most all school projects are "extra work", since students can rarely forge new ground that hasn't been investigated before. Instead, they learn to correctly duplicate known experiments, to get up to SOTA before following their own ideas.
In this case, the machine-vision problem of recognizing blocks with a camera (which inserts distortion and uncertainty to the image) is an interesting challenge on its own.
While viewing a computer screen is easier than most MV problems (the target is self-illuminated, so variable lightsource angles are less of a problem), it could have useful future applications. Especially if someday we're beset by hardened DRM computers that don't allow you to interact programatically- mouse input and LCD output only!
Then the practice of watching a screen with the camera of your PDA might become the only way to take controlled files off those machines.
(Watch the "Ghost in the Shell" movie for bizarre examples of this)
Actually, those people assume no randomness- their "Preview" box shows the entire list of all future pieces that will fall.
(They don't play normal tetris... they also tend to assume that some sides of the pit are unbounded)
WC3 was $59.95 when it came out. (Of course, those were premium "Collector's Editions")
Today, however, it is $22 dollars.