Slashdot Mirror
Are We Searching Google, Or Is Google Searching Us?
An anonymous reader writes "The folks at the Edge have published a short story by George Dyson, Engineer's Dreams. It's a piece that fiction magazines wouldn't publish because it's too technical and technical publications wouldn't print because it's too fictional. It's the story of Google's attempt to map the web turning into something else, something that should interest us. The story contains some interesting observations such as, 'This was the paradox of artificial intelligence: any system simple enough to be understandable will not be complicated enough to behave intelligently; and any system complicated enough to behave intelligently will not be simple enough to understand.' After you read it, you'll be asking the same question the author does — 'Are we searching Google, or is Google searching us?'"
If you're in Russia, Google searches you :)
After you read it, you'll be asking the same question the author does
Do you mean we are supposed to read TFA? Seriously?
Dawkins Revisited: A person is shit's way of making more shit -- Steve Barnett, anthropologist.
Wake me up when it starts teaching the monkeys how to use tools and kill each other. And no Republican jokes, either.
Is Google searching Google some sort of self-discovery process?
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
The best argument against this kind of ridiculous assertion that somehow random information will somehow give rise to intelligence is provided in the old movie Short Circuit. The SAINT 5 robot spends all night reading the encyclopedia and when morning comes, it is suddenly an expert on everything. But its expertise is only in pure knowledge, not the rational use of that knowledge to create something beyond mere identification.
The only way for a robot to grow past its programming is to add the capability to do so. And simply having a system scan data and find correlations isn't going to be enough. There needs to be an action taken on the discovered correlations, and beyond that the actions need to be reprocessed back into the system in a feedback loop. And even further, it is necessary for the program to identify patterns and make intelligent decisions based on those patterns, but the intelligence necessary to make those decisions must come from external sources. I.e. the programmer.
It's a bit outlandish to think that just because a program is constantly watching and processing inputs that it is somehow sentient.
But I have to ask, is it such a bad thing?
You know what it's like, you go to search for something completely innocent and porn comes up. It's not a fault or an idiosyncrasy of the interweb, it's google giving you what you really wanted.
+1 IDisagreeSoHeMustBeATrollOrAnAstroturferOrAShill
I don't know who's searching who, but I do know that I no longer use Google because it's "simply the best". Relevant results are always lost in a torrent of ads, fake review links and e-stores trying to sell me something that's irrelevant.
To the point that I'm not using Google because I genuinely like it any more, but merely because I know the alternatives are even worse. In a few years' time Google went from the best to the lesser evil.
It's... disappointing.
all "magical thinking" in the field of artificial intelligence was reserved for fiction.
There's so much rigorous mathematically described hooey in AI that its hard to tell the naive geniuses from the crackpot morons. Consider this paper by Solomonoff. Brilliant stuff! A fantastic read. Then, at the end, it says:
In our view, however, the most interesting situation in machine learning, arises when we do not know ahead of time what program will solve a given problem and where the machine discovers the program itself. It seems to be very hard to find out much about this by theory alone. Running experiments is crucial.
This is Solomonoff's way of reminding us that he is a mathematician and hasn't actually run any experiments. His other papers make similar pronouncements in the footnotes about the uncomputability of his math or acknowledge the requirement of perfect (aka impractical) training data, etc. He makes it abundantly clear that is work is purely theoretical and unimplementable, but does this stop enthusiastic amateurs from reading his papers and declaring that AI is "solved"? Well no, of course not.
How we know is more important than what we know.
If a system capable of being understood could not act intelligently, then why the hell do we even bother studying the human brain? And further, any attempt at creating artificial intelligence would rely on us not knowing what the hell we are doing?
I am tired of this kind of blanket assumption that anything humans can do that we don't understand or know how to reproduce artificially is somehow incapable of ever being understood or reproduced. We are not so special as to invalidate the existence of the mechanical processes that make us work.
At JavaOne about 3 years ago there was a boffin talk with Gosling, Joy and others and one guy raise the image of hearing something drop through his letter box and then suddenly a little bot appearing in his room with a message "don't worry I'm just indexing your house for Google"
His point was that he had two reactions to this firstly "what a huge invasion of privacy" and second "Great I'll be able to find my car keys".
Of course Google is profiling what people do as they search, indexing everything is what they are about. The question is where this impacts on privacy and what limits we want to put on it.
An Eye for an Eye will make the whole world blind - Gandhi
We had this discussion a little while back. The mythical AI where machines "learn" how to "think" is a long way away or possibly impossible with current technology.
The appearance of intelligence is not intelligence. A recommendations system or search engine may appear intelligent, but the part of the system that processes information "intelligently" was programmed by a person who understood the process. The computer is merely following directions.
Some knowledge based algorithms seem unpredictable when given random data. This is not intelligence either, it is more a result of unintended consequence. You can go back and figure out why it acted a certain way.
It's perfectly possible for insanely complex systems to arise from very simple rules. We cannot grasp the entirety of the system, but we can know exactly how to create it, or perhaps manipulate it.
By way of example: the Mandlebrot set.
Why would anyone engrave "Elbereth"?
We're not searching Google, we're searching the Internet. Google is a tool that can be used (and often is used) to facilitate this search.
Nitpickers are the worst, particularly when they're wrong.
Google searches the internet, but we don't, whilst using it. We search Google, because all the results we want are stored at Google, within Google, and we hopefully find the result we want. Only then are we directed to a site on the internet outside of Google containing the information we searched for.
It is not entirely innaccurate to say that 'We search the internet using Google', but this assumes a logical progession: We search google > because Google searches the internet > so that we cand find what we want on the internet = We used google to search the internet. However, contrary to your misconceptions, it is MORE not LESS accurate to say 'We search Google (to find what we want on the internet).
Turing machines were being assembled into something that was not a Turing machine The author needs a bit of theoretical computer science. However many Turing machines you assemble, you still have a Turing machine.
While doing some debugging on some AJAX work, using tamper data (FF) and Fiddler (IEx) I stumbled upon some nefarious network communications between my mouse* events (over,move,out, click etc.) attached to every single link in googles search results. And there's more! Not only are these events present but they are silently inserted after the page is rendered. Some may say "well this is for older browsers", to that I say, they are not replacing the HREF property on the anchors, they are adding event handlers to mouse* events, and perhaps more that I'm not detecting. You can not see this stuff just by viewing the source. You would need to activate the event that creates the mouse* functions. E.g.: mouse over, and then mouse click gains a new event, so trying to look at the source before the mouse over event occurs yields an null function. Any attempt to look at the source code that google is running (the script handling the events) will be met with a really good obfuscator. Google does this to just about all their public code, e.g.: google maps. The most I can realize about the extra events is that they send a LOT of information to google whenever you click on anything. But don't take my word for it, fire up FF and the latest version of Tamper Data, click 'stop on next line' or whatever engages the debugger (I can't be bothered to look, I'm working on err. something.) and mouse over or click the links on googles search results and watch your data fly over to google, in a rather secretive manner.
It may just be nothing. Every search engine tracks what link you click on, and I think this is one of the more elegant, albeit backwardly incompatible, ways of tracking what links are clicked on. Yahoo does something similar, but they use the 301 permanently moved header with a specially crafted HREF in the anchors, you can see this pretty plainly if you open up yahoo and mouse over the links, they all point to yahoo, then you're redirected to the search. From a coding perspective this is more compatible but annoying to the end user as the link is not what it says it is going to be, it's a yahoo redirector. This means if you try and copy the link from the result you'll get some yahoo bullshit. I like googles method better, but it leaves a lot to be desired in the 'forthcoming' area.
Google also maintains a network of 'adsense' tracker scripts on hundreds of thousands of 3rd party sites, I have several customers that swear by their visitor tracker. It's kinda neat, and it's free, however, I'm sure google does not just ignore the statistics gathered by its tracker. These numerous sites make up a good chunk of the internet, so even if you don't visit google, google sees you indeed. They can track every site that participates, reading referrers and IP addresses, I could imagine some very simple algorithms that could, for the most part, piece together what other non-participating sites you've visited based on the information gathered when you do eventually visit a participating site.
Google Underhandedness IMHO:
1. Adding the even handlers after the page has loaded. There may be a technical reason, but it's just creepy.
2. Sending volumes of information back after each click. There really needs to be a limit. Do you really need my browsing history!?
3. Creating a GPS like grid of sensors on 3rd party sites. This is the creepiest. Google can tell where you are, where you've been and where you're probably going to go with this, and you don't even need to visit google a single time to be added to this network! in fact you don't have any choice whatsoever in the matter!
What Google can do to fix this perception:
1. Quit obfuscating your damn code! It just makes you look guilty when you basically say "Don't look here" in something that is "sneaking" it's way into the source. It's not like google came up with the damn cure for cancer in their JS, what are you try
The real reason is that some of the tubes are bent, and you can't see round corners.
At the bottom of the
Parent is right. As long as there is no way for the programs running on Google's hardware to grow past their original programming (beyond optimization and load-balancing), there will be no Skynet.
Yes, many computer programs work in a feedback loop, and so do all organisms. But as long as only the data entry part of the loop can change, and the system lacks the flexibility to change the type of processing that takes place (the 'program'), no spontaneous evolution will occur.
Several factors are needed to get us to the bleak, dark, machine-vs-human Sci-Fi universe slashdotters know and love.
The first point is the most difficult. It is *not* easy to take pieces out of two programs and build a third program that does things that both do. Whatever OO promises, code is not yet "easy as lego blocks" to assemble. You need very well though-out constraints to mix code in a meaningful way - any self-modifying program would need a small, hard-to-modify kernel that would take care of the mixing mechanism. Nobody knows how to design such a kernel correctly, or what exactly to include as 'genes' (mixable code modules). Computational biology (and biology itself) are hard at work on this problem.
But mixing blocks would not be enough. A successful system would need to build new, unseen blocks by modifying existing ones -- or starting from scratch. How many different things can you say in 20 words? How many of these things make any sort of sense? And how many of those require a very, very specific context to fit into?. The way that evolution can sort this out is by, very slowly, building things that sort-of, kind-of get the job done. However you look at it, there will be huge amounts of trial-and-error involved.
And another problem is that of intelligence "scale". Imagine a super-self-modifying internet worm. The ability to probe and infect does not automatically lead to self-consciousness. There are many, many evolutionary steps from bacteria (very good at self-modification and breeding) to humans. And the current installed base of Internet-connected computers and their "stability" (the time-frame during which a given system remains 'constant') is tiny in comparison to the resources that earths' organisms have had at their disposal for evolutionary purposes. Yes, computers are way fast and this can compensate for some parallelism issues. But I still think that emerging AI is still very, very far off.