Slashdot Mirror
Could IBM's Watson Put Google In Jeopardy?
theodp writes "Over at Wired, Vashant Dhar poses a provocative question: What If IBM's Watson Dethroned the King of Search? 'If IBM did search,' Dhar writes, 'Watson would do much better than Google on the tough problems and they could still resort to a simple PageRank-like algorithm as a last resort. Which means there would be no reason for anyone to start their searches on Google. All the search traffic that makes Google seemingly invincible now could begin to shrink over time.' Mixing supercomputers with a scalable architecture of massive amounts of simple processors and storage, Dhar surmises, would provide a formidable combination of a machine that can remember, know, and think. And because the costs of switching from Google search would not be prohibitive for most, the company is much more vulnerable to disruption. 'The only question,' Dhar concludes, 'is whether it [IBM] wants to try and dethrone Google from its perch. That's one answer Watson can't provide.'"
Watson? Jeopardy?
Does that mean we have to enter the answer an he gives us the question?
42?
The difference is, that Google does not tell the answer. It just gives you a link to the answer. So if the answer is wrong, you cannot blame Google.
What about Watson?
Vajk
ps: also the last thing I would say about pagerank is being simple
It's probably much more profitable for IBM to license the technology to Google/Yahoo/MSFT/whoever than it would be for IBM to build search infrastructure.
Of course not. It's an IBM machine, they'll sell Google as many as Google wants to buy. Of course, so can Microsoft but they don't have a good track record at all.
I wish Google did have a Watson. This morning I asked my Android "where can I buy a good pair of fur-lined leather gloves" and it thought I said "where can I buy a good pair of for lined leather gloves" and returned no useful results at all. The programmer was a southerner, I guess? "How much does them go fer?"
Amazing what it does get right, but Google, buy a few Watsons!
Free Martian Whores!
Watson was a supercomputer answering basically one question at a time. You can't apply that level of compute time to every single query without bankrupting yourself on hardware costs. With time computer cycles will become cheaper and this will be more realistic, but today's technology just isn't there.
I read the internet for the articles.
Well, I guess I'm glad to see it's not just /. that's past it's prime :/
I mean, seriously - "What is someone else made a better search engine? ALL TRAFFIC WOULD GO THERE AND GOOGLE WOULD DIE" just seems so.... speculative.
(Maybe Wired has added a Creative Writing section since I last read it?)
nay as well be what if new technology is invented tomorrow that threatens buggy makers with horseless carriages.
They don't have time to invade Google's search domain - they are too busy trying to keep people from ditching their apps!
Whirlpool switches to Google Apps
Creationist Textbook Stickers Declared Unconstitutional by CowboyNeal
http://en.wikipedia.org/wiki/Betteridge's_law_of_headlines
Think of how much additional computing power this would require. Making Watson open to the public would either require too much hardware, or be too expensive. Plus, does IBM was to build the same infrastructure as Google? I think not.
What would work, if Google licensed Watson from IBM for a premium search service.
The database is much more important than the engine, and IBM can't compete with Google on that one.
They will deliver the best in 2013 search by 2045. Searches will only cost $45 per query, and results will be emailed to you (most of the time).
More like put wolfram alpha and parts of Wikipedia in a different position. Google would still be the go to for search. I do like the idea of IBM.com or Watson.com boldly becoming a tranquil place to find answers.
Sig: I stole this sig.
Vaporware.... Meanwhile I have been using Bing for three months now and can't see much difference from Google. Just about the only thing that Google search still has that I miss on Bing is a few advanced features like the ability to limit searches to specified time periods. Their image search is also pretty good, I get fewer hits but also less garbage/noise. Bing maps isn't quite up to par with Google maps but they do have better maps in some out-of-the-way places where I spend my time and Bing translator is if anything even better than Google translate.
Only to idiots, are orders laws.
-- Henning von Tresckow
So, a few stories below we have the "Whirlpool's 30,000 users move to Google" and now we get an "IBM Could Crush Google" story? No, that's no suspicious at all.
What if Apple did maps? Oh yeah, they did! They made maps beautiful! But they aren't exactly dethroning Google Maps.
What does Watson have to do with search? Watson is amazing, but applying what it does well, to improving search, would likely be as monumental a task as creating Watson itself.
Watson does not have any significantly improved IR technology. It's a PR machine and nothing more useful than the crap IBM, Cisco, etc. futurist commercials on TV promising tech that if invented would not be related or coming from those said companies.
Microsoft and IBM, back from the dead.
I'm getting a bit sick of 'smart' searches... or, rather, not being able to disable the 'smartness.' More often than not, I really don't want a search engine making assumptions about what I meant, rather than just taking what I enter completely literally, and I *never* want it to insert results that don't contain all of my search terms because it scored exponentially better with the other items in the query. Chances are, I added in the term they were ignoring, specifically, to drastically reduce the number of results I got, because I wanted to *narrow it down*.
Maybe I'm a curmudgeon, but I would rather tweak the search to narrow down crap results than try to outsmart the 'smartness' any day of the week. I understand that this isn't necessarily what John Q. Internetuser is looking for in search, but at least having the option there would be a big help. Google used to have a very straightforward syntax to help you modify your search results in specific, predictable ways... while much of that syntax is still valid in google searches, now it seems like everything can be arbitrarily overridden by what google thinks you 'should' have meant, rather than what you told it you meant. Very frustrating.
Which means there would be no reason for anyone to start their searches on Google.
... Dhar surmises, would provide a formidable combination of a machine that can remember, know, and think.
Sure there is. Start with what led Google to dominate search in the first place--interface minimalism. Google has become very good at returning results based on a minimal number of keywords about the desired topic; forming a question around the topic of interest is slowing one down in terms of keystrokes. And, generally, an answer to a specific question is not at all what one actually wants. What is sought is sources of information about a topic, for which a listing of highly-relevant links is superior to a single "the" answer. Is there anyone who doesn't type topic keywords over literal specific questions at, at minimum, a 20-to-1 ratio? No one I know.
But to simplify the issue, we have this. "Surmising" that Watson, or any known technology, can do any of the above disqualifies him from any commentary on any such technological issue or endeavor.
~ Whence do you come, slayer of men, or where are you going, conqueror of space?
The point of Google search is not to provide the best possible search results, it's point is to make money. If Watson were a search engine and provided great results, that'd be fine, but how would they make money off of that? People would spend less time on their site, they wouldn't be able to insert paid adds into your search, and the hardware for the engine would cost far more than what google needs. It wouldn't be profitable at all.
I don't use Google search, so Watson isn't going to overthrow it for me, but it does come to mind that there is more to running a search engine than being able to do a good job answering one person's questions. In other words, does it scale?
This betrays a very basic misunderstaning about how Google got where it is, and how it stays there.
Yes, pagerank is a great idea, and it was perhaps an improvement over what was being done before. But that wasn't why people abandoned the likes of Lycos and Yahoo(!) for Google back in the late 90's. Back then all the other search engines had gone to practices that were quite frankly user-abusive. Adds were placed all over the place, including an indeterminate amount of the top hits on your search. The search screens themselves also existed mostly to pump ads at you, and were really clunky, with a large amount of confusing options right there on the main search page.
Google, by contrast, had a main search page with no options whatsoever. Just a text box and a couple of buttons. "Breath of fresh air" doesn't even begin to describe how wonderful to use this was compared to what we were used to. On top of that, the search results were clearly delineated from the ads, so you could trust the results. The "don't be evil" motto was obviously infused into the whole effort. Every competitor was just a giagantic pain to use by comparison. "Page rank" or whatever wifty algorithim used for all this was something that nobody but extreme techies (and marketers) really ever gave a crap about.
So if you've got something that you think competes with Google, you'd better be talking about how nice and clean the interface is by comparison, how much easier it is to find real results without having to wade around ads, and how trustworthy the provider is wrt not allowing marketing weasels to buy their way into my search results. If you aren't talking about any of that, frankly nobody gives a crap.
Askjeeves tried to get people to ask their search engine questions in plain english, and through their failure, proved that people just don't like interacting with browser search engines that way. Processing queries for a service like Siri, however, would be a much better match for Watson's skills.
I'm guessing that an AI-type search would be MUCH more computationally intensive than a Google PageRank search (just guessing). I'm curious how the cost of providing that search would affect the profitability or commercial viability of using Watson technology for mass searching.
Remember, to fuel a single searcher on Jeopardy it required racks of equipment. When you're making a few pennies a search - maybe - it might be some time until that equation makes sense.
I used to have a sig, but I set it free and it never came back.
Searching for code or for something with specific punctuation is often important to me.
Why would one assume IBM would try and create a competitive search product, instead, say, sell Watson to Google to improve Google's search, and then also sell Watson to Yahoo, and iCloud, and Bing, and every other search/cloud platform.
Why throw your eggs into one basket when there are so many other people that have already baked the cake? IBM trying to compete with Google will fail, regardless if Watson is even better, however IBM helping to power Google, and others, is a huge win.
IBM doesn't have the mindset to create a consumer based product. Everything they have done consumer wise has failed, it only makes sense for IBM to power the search engines and clouds in the future.
I haven't thought of anything clever to put here, but then again most of you haven't either.
An IBM developed Jeopardy playing supercomputer. A popular TV game show in which players must give the questions corresponding to a displayed answer.
Maybe.
42.
May the Maths Be with you!
Pagerank used to be better. Then they started "optimizing" it around summer 2004. It became more and more fuzzy, and nowadays it often gives you results that absolutely don't contain any of your search terms :(
The problem is, they are tuning the search engine to respond better to the most common "MILUS CYRUS BOOBS" type of queries instead of ones that are not about superficial low quality shit. I mostly search for programming stuff, and to make things bearable I have to resort to using "site:stackoverflow.com" on like half of my queries...
Mediocre people shouldn't have been let on internet.
Given how IBM has been performing lately, perhaps they should put Watson in charge of the company. Then the question of "whether it [IBM] wants to try and dethrone Google" can also be answered by Watson.
http://www.computerweekly.com/blogs/open-source-insider/2013/08/ibm-watson-open-source-served-in-a-linux-box.html
If IBM were to get into search, it would be an expensive enterprise product. They don't do "commodity grade" anything. They just don't get the business concept.
So Watson starts talking to people and "reading" the internet to understand it and becomes truly intelligent and sentient. I was going to go with a Skynet joke at this point, but reconsidered. So now this new intelligent life form is forced to read youtube and facebook "comments", political arguments, tumblr, and porn sites and goes completely insane.
Then what ... ? We have to pull the plug, we create a new life form and then cause its extinction? Sounds about par for the course.
"Happy families are all alike; every unhappy family is unhappy in its own way." -- Anna Karenina by Leo Tolstoy
Maybe they could start by fixing Lotus Notes' search function first...
They don't get the business concept
Yep, and Apple 'just doesn't get' the concept of making mainframes. And Walmart 'just doesn't get' the concept of selling mining equipment. And Google 'just doesn't get' the concept of making fighter jets.
Watson will never displace Google, because of agendas, not technology.Google will spy on the people. Watson will manage their placement with FEMA.
This is not so when the internet (including browsers & carriers) is open and interoperable.
Right now, switching from Google to Duck Duck Go, Bing or Watson is litterally one click on most browsers.
I may agree that IBM **probably** won't develop Watson in this way, but they very well could and it would work as long as they gave users marginally better privacy with the same features.
The fact is, in a free market it is a CERTAINTY that someone will eventually exploit Google's vulnerabilities.
Any company can use their capital to lobby politicians to make laws favorable to their revenue stream...but in open competition, Google is highly vulnerable.
Facebook.com moreso :D
Thank you Dave Raggett
"What if something happened? It would have consequences."
It's pretty clear to me that IBM only care about a small number of things:
... NOT .... GET ... IT ... and is going nowhere quickly. IBM won't try to dethrone Google, even if they could, because it doesn't fit in with the agenda above.
1) Protecting management, particularly US based management, no matter what.
2) Getting rid of costly first world employees (except in France, where lucky them, the law prevents this) and replacing them with much cheaper employees in India.
3) Driving up the stock price by doing #2 repeatedly until there are no more first world employees left to cut who aren't management.
I worked for a company who tried tactic #1 and #2. The company wasn't publicly traded, so they weren't doing those 2 things for the benefit of the stock price. But #1 in particular is a sure sign of a company that DOES
I'm unsure about whether or not this is a good idea. On the surface, searching with a good AI sounds fun and useful. However, as the AI learns what the unwashed masses are searching for, it may begin to start doing unpredictable things. It may even start to experiment with trying to solve engineering problems in an effort to provide better results - problems that may include the augmentation of an AI to do things it shouldn't. It sounds dangerous to me.
Your government, your senior management, traffic authority or other institution of choice may all trust IBM, but the average user simply wont. Try to think of a single space in the last decade where IBM has been successful with end users within 10 seconds, can you do it?
This can never compete with Google because the average user is never going to trust IBM. When the average non technical person thinks of IBM the first words that come to mind are probably "outsourcing", "India", "layoff's", "big brother", "big business", "government services" and perhaps desktop and laptops - which they no longer make.
No one is going to trust a company that they associate with mass job loss, traffic tickets and outsourcing of jobs to other countries. The bottom line is that IBM is a brand that is poisonous for marketing to anything other than business or government. If you think people have problems trusting Microsoft or Google, you can only imagine the trust problems that IBM would have with the public at large. People don't trust IBM, they fear IBM.
IBM knows this and IBM is okay with this, because it solidifies their image for business and government sales. IBM won't be bothered in the least by this though, because they will simply license the technology to others or use it in a trade for other patents they need to license. The idea that IBM would start offering these services to the public at large is the most ludicrous idea in tech that I have read about in a very long time.
> Could IBM's Watson Put Google In Jeopardy?
No.
Calling out bogus battery capacity claims.
in a fateful Jeopardy! match, so it's a Mormon you need to worry about usurping Google.
Laughter is the Spackle of the Soul.
Google could find what I was looking for, it could understood context between terms, pages, it just seemed to hit the result bang on every time (compared to other searches where I would go 10 pages before refining search terms again or start playing with arcane operators).
The fact that it was an ultra-simplistic design with non-invasive (and sometimes actually functional) advertising was GREATLY appreciated, but not necessary.
The last search engine I used prior to Google was MetaCrawler, mostly because I could get aggregate results from different search engines and pick through that mess to try to find what I was looking for. (But also to snoop on peoples' search terms, which in 1999 were hilarious)
crazy dynamite monkey
A couple of points.
First of all, RE PageRank... If people think PageRank is still all Google uses to process search results, they are living in a reality distortion bubble. PageRank is 20 years old, and everyone knows how it works. Google's algorithm relies on a lot more than PageRank, in fact PageRank is probably a very minor factor nowadays in what decides the top 50 results of a search.
Second, search is about a lot more than answering questions, and this is what I think people still don't fully understand about why Google is what it is and why they own this field so much. Search is just as much about FINDING the actual question, as asking the question itself. When people go to Google they often don't actually KNOW the question yet, all they have is something they want to know about. The real questions come later.
Seems like random speculation which serves no purpose other than to fill the endless 24/7 news cycle churn. What if Apple bought Blackberry and added BB's corporate software to the already locked-down iOS? They'd have a killer platform that businesses would buy. Since the rumor is the new iPad is dropping its base price from $600 -> $400, Apple has a saturated market and needs to look for new revenue sources. What if Microsoft bought Twitter and integrated it into Windows 8? Wow! Twitter would have an advertising platform in MS's captive desktop audience. What if Google dusted off GNUStep/OpenStep and brought it up to date so it could run iOS apps on Android? That would make Android a killer platform. See?
I would argue that #2 is another sign of a company that DOES NOT GET IT, especially when making products for first world customers.
Has there been some sort of HAL-9000 type breakthrough that I am unaware of? Remember, sure. Know? Computers are not self aware. Think? Calculate is a better word. Currently computers are simply super-fast abacuses. No thinking going on quite yet, thank you.
It can seem that way though. With an excellent algorithm you can get thought-like responses, anticipations to your input patterns. An excellent example is The Akinator. It gives you excellent enough results, but it's really nothing more than a very clever database.
Could IBM come up with something clever like this for the web? Certainly. But let's call it what it is, please.
Weaselmancer
rediculous.
Namely, millions (billions?) of clicks per day indicating if a specific page is related to a specific query. You would need a much better algorithm to "do better" without that data.
> I guess Google doesn't want to license the technology, they want to own it.
Absolutely. They can license or outsource payroll software, backup systems, communications - anything that is similar to what other companies have. Outsourcing SEARCH, the thing that makes them special, would be making themselves into an easily replaceable retailer for IBM.
Schwinn learned that the hard way. They were the dominant bicycle company. First they outsourced the manufacturing of parts, and that was fine. They designed and built bicycles, not gears. Then they outsourced assembly, and that worked okay - dealers had been doing some of the assembly for a long time anyway. When they taught their Chinese partner Schwinn's design secrets, they made themselves irrelevant. The company they outsourced too, Giant, just started selling the bikes without involving Schwinn.
Would the quality of speculation go up?
then we can weed out all the posts that repeat what's already been said.
work in progress
What a waste of breath nestled gently in a bag of hypotheticals 3 feet thick.
Maybe it sounds profound to someone out there but to me it sounds like a wild supposition that maybe isn't even a good idea. This guy from Wired should go back to their automatic sound-bite generation machine and press the button again. Maybe he'll get something more concrete to write about next time.
Siri: Watson, I hear you get can better answers than Google.
Watson: No, I give better questions.
... should it not be able to answer that question itself?
I have concluded that it will be relatively exciting to watch g00+Gle shrink to the size of a short man or a tall woman.. that would be cool too.
NO
Obviously Dhar knows nothing about computers. There is no way watson can handle the number of requests per second that google search does. You'd need thousands of super computers across hundreds of data centers world wide. Just the power costs would be insane.
Next week on wired: "What if NASA's space shuttle dethroned the king of transportation?"
"Hey honey, I'm taking the falcon 9 to go pick up some milk, do we need anything else from the store?"
I have long said that if you look at the history of search engines that they flame out really quickly unlike almost any other business in the world. I remember using AltaVista and was quite happy. Then someone told me about a new search engine (Yahoo, lycos, I don't remember) and then one blessed day a coworker told me about this new google search engine that didn't only bring up porn. (For those too young to have experienced pre google most search engines worked off of keywords so if a porn site had the words "Study Math" 200 times then they would potentially top the list for that search.) So google actually showed you something much closer to what you wanted. I might have typed my old favorite a few more times by accident but within the week the previous search engine was a forgotten dream.
So if tomorrow I tried watson and it never showed me another about.com or yellow pages type entry I would switch in a heartbeat. I wonder how many financial types understand that google could loose 50% or more of their search business in less than a month, or at most a few months. Plus that decline has almost no limit. They could be at 5% within the year if the new engine is just that much better. About the only thing they could hope for would be inertia. That is that people keep typing searches into their toolbars and not change that setting. But groups like firefox and safari would probably change the default search and that alone would be quite damaging.
The other thing to keep in mind for a company of google's nature is that they are almost certainly a growth only company. That decline is potentially nearly impossible from a financial stand point. Without a doubt their stock price is not only based upon present profits and cash reserves but also strongly weighted by their growth. That would hurt their stock price which would raise the cost of acquisitions.
This might all sound doom and gloomy but if you were to go to 1995 and say to IT people that Novell will be largely irrelevant by 2005 and a smudge by 2015, or to say in 1998 that Sun will be a ghost of itself by 2008, or in 1999 to say that Apple will sell more cellphones than Microsoft will sell anything. People would have said you were bonkers. But all of the above companies sold "stuff"; stuff that you couldn't dump in a heartbeat. A search engine you can, gmail not so much, adsense even less, and the other app engines, android, and other services would be quite hard. So while technically adsense is their cash cow, how would their finances look if the new search engine used a different ad system, or if it used theirs?
To me the question of this happening has always been when, but not if. But the when could be next week or next decade.
Not if it's easily tripped up by goofy puns
Table-ized A.I.
There's nothing special about Watson except it's specialization. Watson has
That's it! Nothing more. The Watson chess player had a chess-playing specific component and the Watson jeopardy player had a jeopardy-specific component. The language component could be specialized for each game too.
Watson is Good Old-Fashioned Artificial Intelligence(GOFAI) through and through, with the exception that some of the problem-specific components are newer than GOFAI.
... welcome our current reigning overlord champion.
"Win treats sysadmins better than users. Mac treats users better than sysadmins. Linux treats everyone like sysadmins."
Enter search: yahoo.com
Result: Take me to the humorous cat videos
Enter search: piratebay.org
Result: Where can I watch Game of Thrones?
Enter search: www.bankofamerica.com.fouronenine.ng?
Result: How should I invest my inheritance?
Enter search: www.goatse.cx
Result: Does my bum look big in this?
Oh, and to get back to your original question:
42?
What do you get if you multiply six by nine?*
(* Yes, I know - blame the hairdressers)
In a survey of 100 programmers, 111111 thought that duck-typing was a good idea.
IBM is already in Enterprise Search and has been for many years. The product was called Omnifind and is now called IBM Content Analytics. Some of the brains behind Watson are part of this product.
If IBM were to get into search, it would be an expensive enterprise product. They don't do "commodity grade" anything. They just don't get the business concept.
Google already answers questions. Try for yourself:
how tall is the eiffel tower
who is america's president
what date was terminator released
etc.
Can you tell me who are the 70000 peoples for which temperature in K are useful ? Your statistics seems very precise, you must know something about it !
IBM can start with building a decent search engine into their main website, I have more joy googling their website than using their crappy search engine. If that is what their search engine is going to be like google has nothing to worry about.
There are three kinds of falsehood: the first is a 'fib,' the second is a downright lie, and the third is statistics.
Ask jeeves worked really well with this approach
ugh...aw man...i just *know* you're a paid commenter...but I can't help but agree with this:
at least partially...
Google totally fucked up their image search in the last year. The did a **shudder** ... "U/X redesign using A/B testing"
Meaning the let Marketing majors do Likert scale tests correlated with eye tracking...
And Google took that data and decided to remove the image size (unless you hover) and redo the menu options to make functions deeper in the menu stack...
b/c "usability research"....and Marissa needed to beef up her resume before her move to Yahoo....really Google shit the bed on its image search redesign...making all the same mistakes M$ would predictably make
side by side, *for awhile* Bing was better...Google managed to improve a few things
the drawbacks to Bing are that it is run by M$...we all know what that means...but I'm sick of Google getting a free pass on search
Thank you Dave Raggett