The Future of Search

should have put that in context. Yeah. Yeah, it turns out that without context, it can be a really confusing thing for me, also for computers. Because you know, today we wanted to talk a little bit about the future of search on the Internet. And the reason why we're bringing this up is the week that we're recording this podcast. The today's date as we record is May eighteen. Now, this past week, there was some news that came out from Microsoft and from Google about some additions to their

their search products. So Bing and Google obviously those would be the two UM and uh. And in both cases there's some information about kind of a new approach to search, and not not even that new, it's just that there's

sort of a new way of presenting the information. But it's sort of a kind of a trend that we're seeing building towards something we call semantic search, which is sort of an intelligent search that's pulling information it's relevant to your search query from multiple sources and putting it together in a way that's understandable for you so that you don't have to actually do all that work yourself.

Where you you know, when you typically do a search, you get all the different results, and then you have to go and visit the different sites find out if the information you need is there. Maybe it has of the info you need, but not everything, So you have to go to another link and see if maybe that

has twenty of the fort that was missing. And so you're piecing it all together bit by bit, and you're having to make incredible notes and if this is really data, and and so you might even have an enormous spreadsheet right and it just it's a lot of work. Well, the idea behind semantic search is a lot of that work gets done for you, and you look for something very specific and you'll get very specific results back, even

if those results are pulled from multiple sources. So in other words, you don't have to do all that work. There's still a lot of work that has to be done, though, So we're gonna cover all of that. And I think the best way to start is when we talk about how typically a a search is done in the traditional it's kind of weird to say traditional when you're talking about the Internet, but traditional Internet search style. Yeah, yeah,

well you think about it. There there have been search engines out since probably around um, the early early nineties really, and that that was when we have been talking about the public using the Internet as tool for information, a tool for communications and entertainment really um. And these the earliest search engine relied on some some pretty brute force technology, one of which was was either a a a script called a spider that would go crawling the web, you

get it. It would it would going here, yeah, it would go link to link. It would actually follow links and index pages. It would it would record the content on pages. Now it did not understand that content, but just merely said all right, these are the words that appear on this particular web page, and so that gets indexed. And then it would go to the next link and it would index that one. And you know, there's what I'm saying, one spider. Clearly, that are countless spiders crawling

the web, and there still are. Yeah, absolutely absolutely, And a lot of the search engines, the ones that that we use today, still use that technology, but they're not necessarily as reliant on that technology alone as they used to be. There's more sophistication on the back end. Yeah. And then you have Yahoo, which used a very uh It's genesis used a very um, I don't know, would you say, a intelligent way of going about indexing web content. Out of all the web indexes in the world, Who's

was the most Yes, exactly. They had people. It's a little star trek to reference, going going through and recording what was on web pages. Um, which is again you think about it, that's an excellent way to do it. It's also very time consuming and very expensive pay people to do this, right, it turns out people will not do this for free, all the kindness of their heart.

And with quadrillions of web pages out there these days. Yeah, back in the early days of the web, even then, indexing everything was a was a huge effort, all right, it was. I mean it's not like it's not like they were man, I remember back when there were only three web pages, so easy. I was like, how do I find one? Oh? Right, I need to go to this one of the three. Now it was always a little more complex than that. And so by the time

that you know the the we get to Yahoo coming around. Uh, it's it's a big job going through and checking all these pages. And of course it gets even more complex today when we have dynamic web pages. Because back in the old days, when you would create a web page, it was a very static document. Yeah, it was written in in hypertext markup language STML, and for the most

part there you know, there weren't scripts in there. There weren't server generated pages like you'd find with something like PHP, for example, and people just didn't make that many changes like you might you might make a reference page for something, so it's almost like an article in a magazine or a page out of a book. It did not change

very much once you put it up there. Now, of course you would always have the the the required under a construction gift in the bottom to let people know that, hey, there's still more stuff I'm going to add to this, so you should come back again, yes, usually accompanied by a MIDI file. Also, there was very I can't believe we've never been this. When we talk about the stereotypical stuff you would find on the old web pages, there'd

be a visitor counter. Oh I forgot about the little It would look like an old analog digit counter, and it would go up each type someone would visit it, and there were people who were really proud of their visitor numbers and other people who are very upset that they would be like hovering around seven and you know six of those are your mom? So, um, yeah, what not the it's the universal mom. Sorry sorry but but yeah. So. So the typical way is that these spiders crawl out

across the web. They index the pages. That index is built into the search engine, so that when you type in a query, the search engine compares your query against the database and it starts to try and pull the pages that appear to correspond with whatever your query is. Now, in the early days, this was very primitive, where essentially you were just getting pages ranked, sometimes just by the number of times a certain term might have here on

the page. Oh yeah, they used to uh the early days of search engine optimization UM, which is basically if if you're unfamiliar with the term a way to get more people to visit your website, so they would mention certain words over and over again. So it's a law firm, where where are lawyers and attorneys are? Blah blah blah.

And if you have a legal case that needs lawyers and attorneys at this law firm, this law firm can provide you lawyers and attorneys for your lawsuit, you know, etcetera, etcetera. And then you would even find on some pages there would be like at the very bottom of the page all this tiny, tiny, tiny tiny tech sometimes in the same color as the background of the page, so you'd have to highlight to see it, and all that would

be would be a list of unrelated terms. And it was all about trying to game the system so that the web search engine would point you at that page, because the web search engine wasn't smart. It didn't know that that page wasn't really about kayaking, Right, you type in kayaking in this page pops up and you go there and instead of it being a page about kayaking, it's something else, like a car dealership, And you're like, what the heck is this? Why is this on kayaking

right now. I guess I could fit a kayak on some of these cars, but other than that, I don't really get it. And it turns out they were trying to just drive more traffic, which most of us. You know, you sit there and you think about for two seconds, you think that's a stupid way to drive traffic, because if if you go to a page that does not have the stuff you're looking for, you are not going

to stand that page for very long. But in the early days of the web, people would do practically anything to try and get you to their website and rename photos so that they would work on search terms. The kayak page, the kayak photo one. Also my favorite, which is you go to bookmark a page and you realize that the page was titled something about a paragraph long with just search terms. Yeah, yeah, which just drove me crazy.

So at any rate, search engines have become much more sophisticated since then, where they are able to weed out a lot of that stuff. Google, for instance, it's very famous. They have their algorithm and the Google algorithm page rank. Yeah, the page rank is a lot different than just trying to find a page that has a certain term A certain number of times. In that case, page rank also takes into account how many other sites link into that page.

So if a lot of sites link into that page, the Google algorithm draws the conclusion that that page must be good because a lot of people are linking to it, and if they, if there are a lot of links that point to that page, then by Google's definition, that one must be more relevant than one that has fewer links into it. Now we all know that doesn't necessarily correspond with reality, but in general it's a fairly safe approach.

And another thing it does is uh it recognized as attempts to try to game the system, like we were just talking about. If you do the lawyers and attorneys law firm thing that Jonathan was just talking about, it will go wait a minute, you're obviously over linking. You know, you're over quoting these terms, and I'm going to rank you lower because you are obviously spamming essentially page ten or eleven of the search results. And most people don't

go much further beyond the first page. Some people do

the first two or three. I because I do research for a living, I will often dig fairly deep down to try and find especially if I find something that looks really uh like a remarkable fact, I kind of want to make sure I get some confirmation that this, in fact is true, and I want to see if I can find anything that does not reference the same source material as whatever I found, because you know, you want to you want confirmation, but you don't want to

just like look at eight articles that are all based upon the same news press release, right, because then you're just like, well, all this is regurgitation of the same facts. I need something. I need an independent confirmation that this is true. So yeah, there's you know, but for most people, they don't tend to go beyond the first couple of pages. Um, but this is all still ultimately kind of a dumb search. Yeah,

go ahead, I'm sorry. Well, I was just gonna say that Larry Page and Sarah G. Brand when they launched Google they revolutionized search because it made it so much more effective than it had been in the past, and it really weeded out a lot of the people who couldn't do something as sophisticated, uh search the web as sophisticated as you know, in such a sophisticated way as

Google was doing it. Now. Of course, Microsoft came out with Bing, which uh, if you pull out your fanboy is um, you know, and put it aside for a minute. From whichever of the major tech companies you prefer. I think Bing and Google both do an excellent job of searching the web. They do. You'll come up with different results when you go between the two, but they'll both Um, they'll both come up with excellent results, which is why I like some search engines like Dogpile, which aggregate search

results from these different search engines. Um. But but yeah, I mean they these guys are obviously the heavy hitters because they're very good at coming up with results that people find effective. And that's why they're the big, the big guys in the market, if you will. Now, but they're not stopping and resting on their laurels either. Right, Well, we've pretty much covered what the present day search uh landscape is. Now let's talk a little bit about the future.

And some of this is rolling out presently. In fact, by the time this podcast goes live, many of you may have access to the tools we're gonna be talking about because Google, for example, is rolling out what they call the knowledge graph and being has this nose as in k in Ows approach, where it has three different columns called search knows, being knows, and friends know and um uh. In both cases, they're trying to incorporate a little bit of what we think of as a semantic search. Now.

Semantic search is where the search engine has at some level a comprehension of what it is you're trying to find. And that's really important because, as it turns out, people like to name stuff similar things, even if those two things are themselves not similar. I'll give you a very easy example. Let's say you're doing a web search on son s u N. Well, that could mean the Sun, as in the star that's at the center of our solar system, or it could mean something like Sun Microsystems.

And without any other context the web, the search engine doesn't know what you want. So it's going to start pulling up stuff that ranks really high based upon whatever algorithm they use, whether it's Googles or Microsoft's or whatever search engine, and it's going to present those to you on a page and and maybe mixed up it. Maybe that you've got, you know, three or four astronomical links

and three or four technology links. And that's fine, but let's say that you know you have a semantic search where it can tell what it is that you want and every single link on that page is relevant, Well, that would be preferable and there it's tricky to do that, especially if all you're doing is a one word search term, because you know there's not there's no context there. If you added more context in your search, then that would

narrow down the results. So the more terms you put, in general, the more terms you put within a search bar, the narrower the results will be when they're presented to you. But you have to start trying to think the way. You know, whoever is generating the information that you need, you have to kind of think in their way in order to get those search results narrowed properly, because if you put something that's outside of that realm and it makes sense to you, but it didn't necessarily make sense

to them, you might not get a result at all. Um. I think it's important to note too that we're not what we're talking about now is not necessarily the semantic web. Um. The semantic web is uh, and I just wanted we're not gonna get into that because that could be a series of podcasts. UM but basically the idea is similar. What what we're talking about is context, um and uh. In the semantic web, you're trying to create context between two entities, this person doing the web browsing and the

stuff on the other end. Um. And that requires use of a markup language called r d F Resource Description Framework. And uh. There's actually a very excellent article on how stuff works dot com written by pop stuff host co host Tracy Wilson, um and our lovely site director. Um and uh. And if you want to learn more about that, then I would recommend that you go read that article. But um, you know, and at least until we talk

about it. But um, it's not and Google is careful to make this note to They're not this isn't actually the semantic web, but they are trying to create context by referencing what they know about you and the stuff that is out there on the web. So it is it is similar in that sense. Yeah, it's kind of a it's kind of a step towards the entic web.

I think of it as semantic search simply because ultimately it is looking for bits of information about whatever the subject is that you are interested in and presenting it to you in a unified way where you so that you do not have to do all the searching yourself. It will actually get into why that might be a problem too, because there's there's some very legitimate concerns not from a technology standpoint, but a business standpoint about this approach.

But let's say that, Uh, a computer, for example, doesn't know the difference between any two categories of anything without you telling it what that is. So so let's say that you know you've got a picture. Let's say that you've got a computer that can scan things, and it knows, like you can type in what that thing is. All right, so you scan a picture of a puppy dog, and you scan a picture of a locomotive, and you say,

this is a puppy dog, that's a locomotive. Well, the computer doesn't know that a locomotive and a puppy dog don't belong to the same category like other than stuff. It doesn't know that a puppy dog is an animal. Doesn't know a locomotive is a technology, it's a it's a vehicle. Uh. It cannot make that determination until you tell it. And then to get more complicated, let's say William Shakespeare. You say, William Shakespeare is a playwright or

was a playwright, it's not currently working. William Shakespeare was a playwright. William Shakespeare was a man. Well, you also have to let the computer know that not all all people who fit into the man category also fit into the playwright category, right, Nor are all the playwrights men like the computer does not know that automatically. Sorry, I was something like Bacon. The point being that that very small Shakespeare. I'm a mar Lovian myself. Actually it's not true.

I'm a stratford Ian. But the the entire, the entire relationship of all these different categories are things that a computer does not natively understand. So you have to actually build this. We call these things, you know, we call

them taxonomy's, we call them ontologies. It's all about creating these structures so that the computer can relate bits of information to one another in a way that makes sense, because otherwise, when you would do a search term, even using one of these new products, you would just get a a big batch of data that is not organized

in any meaningful way, you know. But this way, when you actually categorize stuff and you build out these these structures, within the computer system, so that computers can understand, all right, we need to put this bit of information here and this bit of information here, and that this is how it's all going to be organized. Then it means something to you when it comes back. So you might have born on, died on, you know, famous works, that kind

of stuff. And in fact, both Bing and Google are working on actually they're rolling out products that incorporate this

sort of result in searches. So with Google it's called the Google Knowledge Graph, and it's a an idea where if you were to do a search on any particular topic, UH, you would have the main search results underneath the search bar, just as as per normal, but on the right there would be a little uh window that could open up that would have information about whatever the subject of your search was. And it's possible that the information you want

is contained within that window. So if you're doing a search on someone famous Leonardo da Vinci, you could end up which is actually an example in one of the videos they show if you do a search for for the Google Knowledge Graph, there's a video and on the Google blog. Let's say you do that that, well, you might get um, famous works of art that Leonardo da Vinci painted as well as some of his contemporaries. Uh. During the Renaissance, he painted his contemporaries. Uh, if they

stood still long enough. He was a very busy man. Anyway, you would get like little details about his life. You would see that kind of stuff, and and maybe the answer to your question could be right there. So you don't have to go into a search result. Uh. Now, if you wanted to know about the inventions Leonardo da Vinci made, and all you're getting back is art, then you would require you to do another search or to actually go to one of the links that pops up

under the normal search results. Um. But the Google says that they have already more than five million objects indexed for this, with three point five billion facts about those objects and about the relationship between objects, so that when you do these searches, you're going to get the most relevant information. And they say that the way that they're determining what is relevant versus not relevant is they're actually

looking at what people are searching for. So if if nine percent of the people who are searching for Leonardo da Vinci are searching for information about his works of art, that's the information that's gonna pop up, and you do that, just do the search without any more modifiers, which makes sense. You know, if the majority of the people are interested in just that one aspect of him, then clearly that's

what should be presented to you. And then anyone who needs more information just goes into the link, just like you would on any other day. Well being is doing a similar thing. Rhyme, rhyme um. They have instead of instead of the two column approaches, because you know, Google's got the search results under the main column, and then to the right it's got the little window that pops up.

Being does three columns. The first column are the search results, so just like a normal search, and then the middle column is the contextual results where it will pull information about whatever it is you're searching for. So let's say you're searching for a city and you you do, um, oh, let's see what would be a good one. Let's or no, not a city. I'm not gonna go say, I'm gonna go to a place that I visited last year that

I loved. Because one of my one of my family members ruled it for a while, Malta, really Malta was ruled by Strickland. Did you not know that? Yes, I took my I got a picture taken in front of my distant relative, and in fact he is part of my branch of the family. So um, they did not give me really awesome seats at restaurants or anything. I was very disappointed in that they should have at least

given you a free Malta milk. I figured I should have at least had a parade, at the very least a parade anyway, but I enjoyed my time on Malta. Let's say, type in Malta and you want to know you know all the different little facts about like where is it located, you know what what kind of climate is it? And the contextual result within being will give you basic information about Malta, and it might even tell you what the weather's like on that day. That kind

of stuff. It's the and this is the stuff that we've seen in search for a while. It's just that now it's organized in such a way that it is pulled out from the regular search results and put it in its own column, so that way it's not interspersed with the links to the different web pages. So it makes it easier to see and then on the third column it's your friends no column and that what that does is it crawls social networking sites like Facebook, Twitter.

Eventually supposedly it will do Google Plus as well, and you know you have to sign into being for this to work. But when you sign in and you've got linked to your accounts, it's going to search through all the people that you're connected to through those accounts and look for anyone who has posted anything about the subject

you've searched for. So, for example, with Malta, if you're a friend of mine and you search for Malta, you might see that it pops up that I've got some pictures of the time when I was in Malta and some some posts about what I did while I was there, and maybe about a restaurant I went to. And that would allow you to do things like touch base with me and say, hey, you know you've been to Malta. I was thinking about going. I wanted to talk to you a little bit about your experience and find out,

you know, what should I do? And uh So it's an idea of making the search more relevant. Now this is really interesting to me because this shows a divergent future of search. Right, You've got the what I call the semantic approach, where's the artificially intelligent approach, where the computer is putting together information based upon algorithms and connections

between objects. And then you have the social approach, where it's pulling information from your friends and and other people you know and giving you sort of the personal approach to search. Two different ways of doing this. I'm curious to see which one's going to win out, and maybe maybe win out is the wrong phrase. I'm curious to see which one people will rely upon more frequently, and it maybe that it all depends on what kind of person you are. I also am sort of dreading this

in a way. The reason I'm dreading it is I'm worried about people doing searches for It's lots of stuff that I post about and then they get results about me, and the next time I log into Facebook, I've got seven and fifty three messages asking me about the restaurants I eat at, the theaters I go to, UH support questions about the multiple things that I've written about. For how stuff works, I mean, you know, you could you could see that for certain people who post a lot

to social networks. It could mean you get a flood of folks contacting you about stuff that you either haven't thought about in a long time or you know, yeah, you just you're like, well, I don't mind sharing that information. It just feels like I'm getting uh buried under requests and UM. I talked about this with some some tech friends of mine, and they all thought I was being a bit of a grouch. But I said, no, seriously,

think about it. Think about all the things you do and post about and so well, isn't that why you post about these things like you post about it in order to share the information with people? I said, sure, But in that case, it's me shooting information out to the world. It's not the world coming to me for information. That's a very different thing. It's it's you know, taking the bullhorn and saying something through that as opposed to having to accept call. After all, after call, people don't

call me, do not call me. Well, there's also another thing to be concerned about for a lot of people, what um. In order to build a context between you and the other web pages out there, it has to know things about you, and some people are shall we say, skittish about uh providing information about themselves. They might be looking for more privacy understandable, and this is one of

those trade offs that that you have to make. Uh. This is incredibly convenient in a lot of ways if you think about it, but in order to do this, you have to give up information about yourself. Otherwise the search engines can't provide that context and the relevancy that you would get with these enhanced search features. So you have to make a decision do you want to share this information with being Do you want to sign in and let them peruse your accounts so that they can

find out that Jonathan knows something about Malta? Or do you want to say, you know what, I don't want to share that information with you, and I will take my chances with the search results. And that's, you know, something that that is an individual, person by person thing to consider, but I think it is worthy of mention because privacy, of course, is one of those things that that is a greater concern for many people. UM and UH.

Another thing to consider is UH the filter bubble, which is a term that Eli Harriser came up with UM or at least he wrote a book about it, and the idea being that the more you search, the more your search engine knows about you, the more stuff you're going to see that is uh, you know, tailored towards your interests, which is okay in way good, But in another way, it's not because you're not seeing uh, divergent opinions.

You're not seeing opinions from people who are not like you. Um, if you are, say a Republican, you're gonna see things that are more conservative and fewer things that are more liberal. Just becomes an echo chamber. Yeah, and and so therefore you are not proposed to opinions from people who are not necessarily you know, don't necessarily hold the same beliefs.

And there are a lot of people. I used politics as an instance because a lot of the the people that I've sort of paid attention to in the political world want to know what's going on on both sides of the aisle. They want to know what people who are like them think, and they want to know what the other people think too, um, so that they can make an informed decision, which I think is you know,

something to be something to be mindful of. Two. Um, not that not that these these search tools are bad in any way, But it's something to consider when you're using them, is that you're trading off privacy to gain this, and you're also UM insulating yourself to just things that you like, which is frankly a very comfortable position. Uh. You know, it's we're staying inside our comfort zone and not and not treading very far outside it. UM. So it's it's you know, those are just a couple of

the things mentioned. Also, I've heard critics say that they don't like UM the idea of of what Google is doing because they said, well, you know, you're you're putting a lot of the information up on from these pages on Google. What incentive do people have to click through to the people you're linking to, Which is where I

was saying. You know, it's not a technological problem, it's a business problem because if you're if your business is web based and you depend very heavily upon ad revenue, then that means you depend upon people visiting your site to get information or to have an experienced whatever your site provides you. You depend upon people coming to that site in order for that impression to be made, and

that impression goes toward UM generating money through advertising. So that because we we assume that if you're on the page that you're seeing the ad or the ad is somehow affecting you, and that that means the advertiser is

getting value for the money they're spending to reach you. Well, if you never go to the site, then and that impression never happens, which means that the the proprietor of that site is out of advertising money, and in fact, the advertisers aren't happy with it either, because it means that their ads aren't being seen. So you know, you've got the advertisers who are unhappy because their ads that

they're creating aren't being seen by anyone. If in fact, you can get all the information you need just through typing in some search terms into Google. Uh, and the the proprietors the sites that actually are providing that information that their owners aren't happy either because no one's going to visit them. So then you think, well, would anyone

even bother to build a website? Um, once you got to a search that was sophisticated enough that all the information comes back upon the search, why would you bother building a site at all? You're really all you're doing at that point is just building a database for Google. That's all it is. It's not you're no longer building something for yourself. You're really just adding to Google's data storage.

So that's a legitimate question. I mean, uh, now, if if the information the way Google is presenting this, it tends to be not a huge amount of information. It's it's stuff that's relevant to your search, and it might answer very basic questions you have about that. But if you want something more thorough than just like a birth date to death date and some major achievements in a in a particular historical person's life, then you would need

to go further in. It's not like you're going to get a full biography with lots of detailed information within that search. Right. What I've what I've seen Google say is that they believe that it will wet people's appetites. Uh, they'll get you started, but if you really want to know more, you need to continue follow the link into the what the other sites and see what else there is?

Um which you know for some of us, um you know, not speaking necessarily for Jonathan, although I suspect it might be the case, uh, you know, for people who are intellectually curious you do start wanting more. It's like, okay, well this is great, uh and it answers my question, but now I want to know more. You kind of fall down the rabbit hole where you just start chasing

more information and uh and offend. That will lead to you looking at related things that had nothing to do with your initial search but are interesting, so you want to pursue them and find out more about them, which is great. Uh, It's just that I know a lot of I don't know. I suspect a lot of advertisers and website owners are a little leery of this. I'm sure. Uh yeah, that's that's one of the one of the issues. And here's another problem. This one is more technological. So

building out these these results, these connections between objects. Like I said, a computer does not natively understand relationships between any two things, so you have to not yet they're working on it, but that means you have to build stuff out. And I wanted to bring something else into

the discussion to kind of give an example. There's another project called the Silk Project, which is a I think it was eleven developers, eleven Dutch developers who put this together, and the idea is that it's a it's sort of a database management system in a way not necessarily meant for the Web, meant more for like an internal system. Like let's say that Chris and I, you know, we

work for how stuff works. Let's say that we build out a um like We're going to build out a wiki that contains all the information about our company and processes within our company, the procedures you're you're supposed to follow, basic information within the company. The Silk Project would allow you to classify the information in different ways. You tag it,

you create meta data. Now, meta data is information about information, and this allows you to to classify the information in ways that the computer can on a very surface level understand understands really the wrong word, but it can work with that classification system. So, for example, if I were to have a whole site about dogs, then you know, dog might be one category. I might even go further up and say it's a mammal and that means that's

an animal, etcetera. But then I would have breeds underneath the dog, and each of those breeds would go back up to ultimately two dog. Right, Well, that would help me build out these relationships between whatever it was that was in my database. So Silk, what you do is you get this tool you classifile the information within your system. So let's say it's lots of different numbers, lot's of different like sales figures and advertising figures, that kind of stuff.

You would classify all that information and then whenever you needed to pull data from your database, you would type in a query into the soap Projects search tool and it would automatically aggregate all the information for you and put together in a way that makes sense and make it easy for you to generate graphs and charts, sometimes automatically, so that you could have a visual way of presenting

that information to someone, so you can make a larger impact. Uh. And this makes a lot of sense because otherwise, again, what you have to do is you have to do searches into all of these individual documents, pull the relevant information, put it into a spreadsheet, manipulate it to the way that you need it, and then create whatever graphics that you want from that. This does it all automatically. However, the first thing you have to do is you have

to tag all that information. If you're building out from scratch, it may not be a big deal. You know, you just it's just part of the building process. You know, you create information and you tag it, and you create a way of tagging it that's consistent across the company.

Because here's another thing. The way Chris thinks and the way I think are not perfectly aligned, and so sometimes he might tag a certain kind of information one way and I might tag it another way, and that might be problematic depending upon how the computer handles tags and classifications.

So that if I were to do a query and Chris were to do a query, and we're doing the same query, but we've both set up our systems in different ways, even though we have the same basic information in both of our systems, so we're pulling from the same pool of data. We've just built out our systems slightly differently, my query might pull up a different result than his because we use two different tagging systems. I think the Silk Project actually tries to take that into account,

but even so, it's a little tricky. So anyway, that's a lot of work for a person to do, because again the computer cannot determine automatically yet what categories certain tags fall into, or certain certain terms fall into not tags. I'm sorry, So without knowing that that means someone else has to go in and tag it themselves. If your database is enormous, then that can be a really long,

work intensive process. And you know, you keep in mind that most companies, most organizations are continually adding to that. That means that you are not just trying to get all the information you already have, you have to do all the incoming information as well. So it can be quite uh an extensive problem. And what you're really doing is your frontloading all that work that you would be doing if you were to do it the old fashioned way. So by frontloading it, you've made the actual handling of

the data further down the line much easier. But it means that those early days are going to be kind of painful, a lot of a lot of data entry stuff. Yeah, well, I think this is it's good to talk about these things. I mean, for most of us, using the search engines

as they have been has been pretty effective. But you know, I do think it is good that Google and Microsoft and and the other companies who are doing search are are not sitting around and and saying, well, you know, we're making billions of dollars as we as we are making search you know, doing this and selling advertising along our our search results. They're actually trying to find ways to make search more efficient and more relevant to those

of us. So I think, um, despite you know, the shortcomings or potential shortcomings depending on on your take on the issues, UM you know, those things, I think have a very real trade off. I don't think that what you're giving up is for nothing. I think, you know, having the ability to have personalized search results can present an amazing positive experience. UM So, so please don't take my my mention of those criticisms as you know, necessarily my opinion on it, because I do have sort of

mixed opinions about about these things. But I do most of all, I think I appreciate that they are trying to find ways to improve um and I think that's always a good thing, especially in technological circles, because these are our tools that can be made um you know, available to thousands and millions more people, um and can make our lives even easier tomorrow than they do today. It's just something that I think we should always take with a grain of salt on what are we giving

up by accepting this? Ultimately, we might reach a point where and navigating the web is like having a personal assistant that is constantly filtering everything for you and pulling stuff for you, so that UM that you have a a really customized um experience, so that when I get on the web and when Chris gets on the web, it will be completely different, even if we're looking for the same information, because it's been tailored to our particular style.

And we'll we'll probably see this rollout to not just you know, things like web searches, but into other products as well. I mean, they're already talking about making sure that this sort of sort of search information is also available mobile platforms. But then you look at things also like Google's Project Glass, the glasses. I can totally see that being part of it too, that the glasses, you know,

linked back to your Google account. Uh, start to take into account how you're using the glasses, and it can help start to customize your experience based upon the way you use a device like that in the real world all the time. And if it's like, Wow, you know, he's really interested in food a lot. So if there's a term he searches for that could possibly be food, remember that because it maybe they's looking for the food and not the place name or character name or whatever.

The answer is food. Food. The answer for me will almost always be food. And I think, uh, I think that pretty much wraps up our discussion. You anything else to add before I wrap it up? No, not really, all right, it was it was kind of fun getting into this because on the surface it seems like a small thing, but when you get into it, there it's it's fascinating. Layers within layers. It's like a great, big

old internet onion. I was gonna say it's like an ogre also that I am going to wrap this up now. I'm gonna go and grab some shwarma, which is it's all all the rage right now thanks to uh, thanks to some some superpowered people. Um, but I was a fan of it before. Hey, you know what, I like shwarma before it was cool. I'm gonna shwarm my hipster. But yeah, I'm gonna go get some food and I'm gonna wrap this up. Guys, if you have any suggestions for future topics that we should talk about in episodes

of this podcast, tech Stuff, let us know. Send us an email our addresses tech Stuff at Discovery dot com or contact us on Facebook or Twitter, or handle it. Both of those is text stuff. Hs W and Chris and I will talk to you again really soon for more on this and thousands of other topics. Is it hastaff works dot com brought to you by the reinvented two thousand twelve camera. It's ready, are you