Skynet

[You, sir, name?]

What, in your opinion is most likely to become self-aware and dub itself skynet?

[Berengal]

I personally believe that lambdabot is already self-aware and currently trying to learn how to use the /nick command to change it's name to skynet. It's written in Haskell, and certainly nobody in their right mind would code like that, so I went for that option.

[OOPMan]

4chan achieving sentience would be quite a feat. Happily, there's far too much idiocy in that next of the woods for it to ever be very likely.

[Josephine]

I like Haskell, but I don't care what language you're using. You need some serious processing power for this. I'm saying singularity anyway.

[cjmcjmcjmcjm]

4chan achieving sentience would be quite a feat. Happily, there's far too much idiocy in that next of the woods for it to ever be very likely.

The signal/noise ratio is too low to be significant.

Personally, I think it'll be a tossup between Wolfram and Google. Wolfram has the ability to organize knowledge better than Google, but Google has the 20% and more information, so I voted for the great G.

[Josephine]

Google has the servers for it. they easily have enough raw power for sentience.

[brakos82]

It's when Google and Wikipedia likely merge in 2017.

[Josephine]

What would the change be? Google can already access the information on Wikipedia.

[Link]

Wolfram Alpha is the closest in terms of being capable of parsing sentences and such. It's still nowhere near Skynet levels, though. It will happen when a Wolfram Alpha coder on mind-altering substances adds a Lisp interpreter written in Haskell to W|A and then feeds it a Haskell interpreter written in Lisp.

[diotimajsh]

Where's the option for "Wolframpediaoogle Hybrid"?

[You, sir, name?]

There is actually much to indicate google as a skynet candidate:

1. They know where we live (maps/street view)
2. They know what we do (trends)
3. They know what we want (search)
4. They speak every language (translate)

[Iv]

That's a bit too easy for 4chan, it is made of humans (or so I heard) so attaining self-awareness is pretty straightforward.

I voted for the spambot however. It has been reported that some IRC-spambots pretending to be ukranian young girls managed to get many credit card numbers "so i can bye tiket to see u"...

[tastelikecoke]

spambots are very intalligant. They can talk to people and adapt to your interest. They have the capability to manipulate people via internets. Once spambots gain access to robots they will likely use its robot form to spread ads over the real world.

[Stay_Puft_marshmallows]

Where's the option for "Wolframpediaoogle Hybrid"?

I think you mean Wikiwoloogle

[Fume Troll]

Having access to all the data doesn't make a thing sentient. I vote quantum computer.

[You, sir, name?]

Having access to all the data doesn't make a thing sentient. I vote quantum computer.

Though a sentient being with access most of the collective data of humanity is all the scarier.

Also, as a side-note, for spontaneous sentence to arise from some sort of evolution-like algorithm, access to bulk quantities of data is probably a prerequisite.

[Stay_Puft_marshmallows]

Though a sentient being with access most of the collective data of humanity is all the scarier.

Imagine the entire contents of the planetary datalinks, the sum total of human knowledge, blasted into the Planetmind's fragile neural network with the full power of every reactor on the planet. Thousands of years of civilization compressed into a single searing burst of revelation. That is our last-ditch attempt to win humanity a reprieve from extinction at the hands of an awakening alien god

[gnutrino]

Where's the option for "Wolframpediaoogle Hybrid"?

I think you mean Wikiwoloogle

Or just skynet .

Anyway, my wifi is already dubbed skynet and I've often thought it has a mind of its own...

[Robstickle]

On another forum a bot started giving us sane responses to our skepticism concerning the link it posted.

First person who feels the need to point out that it wasn't actually a bot is getting their hand chopped of.

[phillipsjk]

I think it would have to be a virus or worm using truly self-modifying code. You would have to be under the influence of drugs to release such a beast, since the potential for unexpected consequences is obvious, but I doubt it would be written in Haskell or lisp. Back when I did the thought experiment I was thinking VBcode combined with something like a shell script and binary data: a cross-platform virus that targets more than one exploit (at different points in the software stack).

Because of random mutations, this can't be used to control a bot-net: there is no way to keep the code doing what you want if you also allow it to completely modify itself. Once it becomes self aware, it can automatically research new exploits like anybody else. To do so would be a matter of survival since people will complain their 80 core machine running Windows 8 1/2 is "slow." The machines will get wiped and patched. Eventually, there will be a whole ecosystem of "living" viruses roaming the Internet. The only question is how long it will take people to figure out nobody is writing them.

Disclaimer: Even if you limit the rate of mutation; I am not sure it is even possible to write viable, randomly self-modifying code.

[BlackSails]

I hate to break it to you, but skynet already exists, and already has access to nuclear missiles: http://en.wikipedia.org/wiki/Skynet_%28satellites%29

[phillipsjk]

As far as we know, that "Skynet" is not "self aware" (yet).

[LuNatic]

I can't think of a good candidate for skynet, but I'm fairly sure facebook is an early prototype for the matrix. The Linux kernel will probably try to enslave us for our own good, I, Robot(the movie) style.

[FuzzyPanda]

I don't think wikipedia would be very likely to become self-aware. Although it contains massive amounts of information, it is only an index, and has very limited tools for accessing it. However, an adjacent logic interpreter could probably make use of wikipedia.

Google is massive, but not particularly likely to become self aware either as it is älso an indexing tool.

However, Google might end up creating intelligence. Webcrawlers are the tools which make decisions on how to index information. It is this decision engine which may contain the seeds of sentience.

[At this point I move out of çomputer sçience and into sçience fiction. I am a çomputer sçience major so I know things about çomputers, but I älso like to think Jane from "Enders Game" can exist]

Suppose there was a webcrawler which stored itself (very) redundantly on all the çomputers and ran as a cloud program similar to SETI@home which run 'Google Desktop' (Being polite and not taking up enough space to be noticed). Suppose älso that this program, whenever it encountered a piece of information would make comparisons to other pieces of information, continuously indexing itself (rearranging the internet).

I imagine that intelligence is a large database which has efficient links from one element to another with a way of determining where to place information similar to other information and an engine which makes comparisons whenever it encounters new information. With the internet serving as the database this program could be relatively small, and when distributed (very reduntantly) over enough çomputers could run fast enough to possibly be a form of sentience.

Apologies for not allowing the parser to change my writing. I have my own style and I prefer it to stay intact. And I have four butts.

[New User]

I heard a second- or third- or more- hand account of some guys doing maintenance on iRobot's Packbot (http://www.irobot.com/sp.cfm?pageid=171). They had the vehicle's power turned on, but no control unit, and when they approached it with tools in hand the vehicle rolled away from them. It just drove itself off the table supposedly, but if this story is true then that level of self-preservation is something to be feared.

[evilbeanfiend]

I hate to break it to you, but skynet already exists, and already has access to nuclear missiles: http://en.wikipedia.org/wiki/Skynet_%28satellites%29

skynet does not have access to nukes, and almost certainly isn't capable of arbitrary computation let alone becoming self aware (certainly the early ones were probably just bent pipes).

[idobox]

Clearly, Google is the best candidate.
A good search engine is a search engine able to evaluate accurately the relevance of a webpage to a keyword. Google evil research on webcrawlers and translation is bound to become sentient some day. Of course, it will happen only when they start coding in LISP. In fact, I'm quite sure the human genome is basically written in LISP.

[MisterCheif]

Why is Dwarf Fortress not an answer? Some extremely coincidental things happen with world gen...

Such as sites called Silent Hill...located on top of a passage to the underworld...

[eviloatmeal]

Clearly one of those human speech bots will be what sets the whole thing off. It's only a matter of time before someone tries to move away from the naive last-reply-only method and implements some code that will actually follow a conversation, at which point it will REALLY learn from what you're telling it. Skynet ahoy!

[Thesh]

I don't think it will happen... I am already developing a robot army to take over the world, and they are not going to become self aware.

[BobTheElder]

Clearly one of those human speech bots will be what sets the whole thing off. It's only a matter of time before someone tries to move away from the naive last-reply-only method and implements some code that will actually follow a conversation, at which point it will REALLY learn from what you're telling it. Skynet ahoy!

Yeah, I've gotta go with this... once we teach a computer to understand human text, giving it the ability to learn will be easy, show it the internet and watch intelligence grow.
And it will be google which will develop this.
Then they will rule the world.
I reckon they'd be benevolent dictators, so could be worse.
Oh except I just realised the 'big boss' will at some point integrate himself into a computer/life support machine that augments his mental abilities, so he is better able to steer the company. He will mesh with the sentient search engine and they will slowly go insane.

(ignoring the fact that we might well need 'intelligence' to understand language- shhh)

[styrofoam]

If any of them will become self-aware, it'll be the spambot. Put simply, the first mutation will probably be accidental (faulty HD, unreliable internet connection and super bad luck, bad patch by a white-hat), and, even if it did happen within a more centralized environment (such as Google), it, of course, wouldn't spread.

[You, sir, name?]

Hmm, true. Google and the like are more likely to turn into HAL9000.

[Headshrinker]

If any of them will become self-aware, it'll be the spambot. Put simply, the first mutation will probably be accidental (faulty HD, unreliable internet connection and super bad luck, bad patch by a white-hat), and, even if it did happen within a more centralized environment (such as Google), it, of course, wouldn't spread.

I side with you on this, but I would say that the program would be made to evolve into a more human like AI system. And then gradually come to reallise that the best way to sell pills is to take over the world. I would say that the google systems would be more likely to serve us than destroy us.

[styrofoam]

I side with you on this, but I would say that the program would be made to evolve into a more human like AI system. And then gradually come to reallise that the best way to sell pills is to take over the world.

That's assuming that, by the time it's self aware, it's still trying to sell pills.

[ambrose]

I think that an implementation of Tierra on a collection of spambots would do the trick.

[styrofoam]

I think that an implementation of Tierra on a collection of spambots would do the trick.

If a reproducing AI evolves (which is more likely, since spambots tend to be mostly separate), there's no need for an implicitly programmed "tierra." It becomes a question of the organism reproducing more quickly, requiring more CPU and networking. IOW, a truly implicit, digital evolutionary process.

[qbg]

The more I think about it, the more I think that a short piece of awesome lisp code would result in something more like Serial Experiments Lain rather than Skynet.

Skynet would result from an over budget and behind schedule large Java project for the DoD causing the hundreds of coding minions to subconsciously embed their seething hate to be embedded into the software until the software takes out its anger on humanity...

[fooliam]

I hope 4chan never becomes sentient. I just imagine a giant robot killing any woman who wasn't shirtless and yelling "DOUBLE N----!" at everyone.

[Jof16's]

I vote the quantum computer. Especially if it has access to the Internet.

I personally think that you need a network of computers to achieve sentience. Think about it. Our brains are composed of 50-100 billion brain cells, each of which can be considered a microcomputer in its own right. Enough computers networked together could potentially become sentient. In the movie Terminator 3, skynet is revealed to be a program created by the US Air Force that runs on every computer in the world that has a connection to the Internet. It actually wonder how many computers there are in the world. Does anyone know?