See Wolfram Alpha in Action: Our Screenshots
Last weekend, we attended a web demo of Wolfram Alpha, a new "computational knowledge engine" based on the work of Stephen Wolfram. Some have dubbed Alpha a "Google killer," but, in reality, it is very different from the standard search engines that we are all familiar with today.
When we got the demo, Wolfram asked us to refrain from publishing any screenshots. Today, however, the Berkman Center posted a video of the public demo Wolfram gave earlier this week, so we think it's only fair that we share our own screenshots with our readers at this point.
Homepage
Query #1: internet users in Europe
Query #2: weather oakland
Query #3: oakland
Query #4: uncle's uncle's brother's son
Query #5: water 550C 3 atm
Query #6: integrate x^3 sin^2 x dx
Query #7: bob
Example of a copy and paste dialog:
Embedding Search Results:
Here is the video of the public demo at the Berkman Center. It is a bit blurry, but it does show Wolfram Alpha in action:
And if you really want a look behind the scenes, here is a look behind the scenes of the Wolfram Alpha datacenter:
Share
Facebook
Comments
Subscribe to comments for this post OR Subscribe to comments for all ReadWriteWeb posts
I think that Google will eventually move into this area over time.
Looks amazing to me. I hope it works with all kind of international locations and specialties. Some experiments before were US-made and therefore usable for US-only.
Can you give us some examples on search results for Asia or Europe? What happens if you search for e.g. chemical formulas?
@stefan - I only got a demo and don't have access to the actual system, but while I don't remember the details, Wolfram briefly showed some info about chemical compounds, but not formulas (I would be very surprised if it wouldn't return some interesting results for this as well).
Except for the language, I don't think Alpha has a strong U.S. bias, so you should be able to get similar information for cities all over the world.
Search is 90% done!!
How on earth could this possibly add any value whatsoever?
This is maybe 1% of the missing 10%!
Pah!
(hint: sarcasm)
Is it just me, or does Wolfram Alpha seem more like a Wikipedia killer than a Google Killer? I mean they both seem to do entirely different jobs. Google helps you find sites on the web, Alpha provides you with an "expert" opinion as to what it thinks the best answer to your question may be.
Just my two cents anyway...
Chris - I agree in some ways: I don't think that it will replace Google. I think that Google answers commercial searches very well - better than this will probably - but Google fails a bit on some informational searches. So if this is as good as it looks then I think that it could steal some market share from Google: you'd use Wolfram for facts and information and Google for commerce and finding sites. At least until Google buys them or builds their own.
Of course informational and factual queries are harder to monetise...
Alpha is your personal genius it isn't Google killer
Already a Fail. If you say "Launching May 2009"... then be open on May 1, 2009.
period.
So, it's sort of like Wikipedia, except that:
1. It cannot answer subjective or political questions, AND
2. Because it's a computer, it will occasionally make farcical comments like "The Statute of Liberty weights 13 oz. and is made out of Dr. Pepper."
Isn't this something of a step backward for search, in that it represents a hierarchical data model rather than the cloud that Google has created?
For instance, when I search for the weather in Oakland now, I get a hundred different sites claiming to give me the weather. But I choose to visit the ones that are the most reliable based on what I know, and based on other people's knowledge (represented by the ranking of the sites on the page), etc.
With this, I'm forced to trust that the computer already made those decisions for me, and is now showing me the most reliable weather forecast. I suppose I could trust it, but it seems backwards to me.
I agree -- it's more like Wikipedia, but probably without the instantaneous updates. Could this be a genius idea without a practical application?
This is objective search results and not subjective search results. Still no solution exists to mesh the two together well enough without human intervention. It wont kill wikipedia, a good portion of it yes, but not all of it.
@Falafulu, @Jason - It isn't a step back from Google -- just a different approach to the problem with different costs/benefits. Google's approach, essentially that of throwing a bunch of statistical analysis at the contents of the web, makes it sort of the Jack of All Trades. It does a pretty decent job at many things, and you can use domain-specific knowledge to help do an even better job at other things.. but what Google's approach lacks is any notion of semantic data modeling and data that constitutes base truth about these semantic entities. Alpha is focused on exactly this blind spot of Google. Rather than providing essentially souped-up text search on the web, it is trying to answer questions about specific topics for which the answer is calculable.
First I thought wolframalpha would be the next generation of google, but now I am a little bit disappointed. It´s something like a very smart wikipedia. It connects data with language. So the site is able to understand my question and find the information i need in the system.
Maybe this is good for technical things, but for all the other things (for which the internet is growing that much in the last years) this isn´t suitable.
What about funny, creepy, interesting, foolish videos? The mashine wouldn´t be able to show something like that, or even webistes with exactly the videos I need.
But not only videos, also satiric textextparts, blogs, peoples opinions on products, multimedia, fun, buisness, nearly everything isn´t suitable for this searching engine.
It´s suites only for technical things, which are 100 procent true und can be prooved. So it makes sure, that the machine isn´t doing any mistakes.
I´m worried that, wolfram alfa will not be able even to be as good as wikipedia.
I had a semi-private (6-8 of us) dinner in Oakland (or Alameda) with Stephen Wolfram way back in 1992. I was a fairly heavy user of Mathematica back then -- mostly for innovation diffusion modeling -- and looked forward to meeting him. Well, I wasn't disappointed. I found him a bit introverted, also modest. When he did speak, his words added much to our collective discourse.
Fast forward and now we have Wolfram Alpha. I don't like the name -- Larry Ellison once noted that at least he didn't name Oracle after himself, a direct reference to Siebel Systems (which Oracle eventually acquired) -- but I like the concept. (Of course, I'm also guilty of being an early Twine proponent, so what can I say.)
"Google killer" stuff is just hype. I doubt that Stephen would use this phrase. But if it's all it seems to be, it will certainly be a useful Google companion, much more so than something like Scirus (which is more like Google Scholar). And there's no reason why Google couldn't acquire Wolfram Alpha. Hey, maybe Oracle could acquire Wolfram Alpha; the Oracle M&A team has been pretty much on target for the past several years. Wouldn't this be a interesting play ... and it would help make the brand known to a much wider audience. Maybe Microsoft would want it, but they'd probably screw it up. And using a stretch of the Oracle rationale, Cisco or EMC might want it.
Anyway, it's good to see progress when it comes to search. It's good for Google to have commercial competition, which it really doesn't these days. I wish Wolfram Alpha all the best -- and hope all of us get a cool search appliance.
(To Stephen, just in case you read this. At our dinner I was the guest of a beautiful blonde named "Susan." I don't recall her last name, and it wasn't her maiden name, but the name she took from her first marriage. Since she was a beauty in her own right -- and the most beautiful egghead I've ever seen -- you probably remember her. She eventually got remarried to a guy whom she met on a scuba diving trip. At that time I was in the invasive medical devices and biodiagnostics sector, early bioMEMS stuff. Our precision engineering and flexible automation company made the equipment for manufacturing balloon catheters, atherectomy devices, RF ablation probes, in vivo blood gas and pressure sensors, and various biodiagnostic, mostly in vitro, rapid test kits. After that I went to Samsung, Microsoft, Oracle, META Group, in this order. And my day job these days is running the U.S. and China ops for a Spanish renewables company in solar [standard modules, LCPV, MCPV], wind [small wind] and electric vehicles [FCVs]; I'm primarily based out of our China HQ in Qingdao and have been living in China for the past five years.)
Is it a mythical creature that is part wolf, part ram, part alpha? Does it devour on google’s when its hungry? When it eats aol’s for breakfast does it do what most of us do and just regurgitate it back out of our system? Is yahoo its desert or salad! Probably why I don’t like desert or salad for that matter.These are awesome video's if you would like to check out the entire presentation I have it blogged at http://askwolframalpha.com
1. It cannot answer subjective or political questions, AND
2. Because it's a computer, it will occasionally make farcical comments like "The Statute of Liberty weights 13 oz. and is made out of Dr. Peppe
All I can say is bring it on. Bring anything on that doesn't reduce the written word to a bunch of nonsense. I almost always dismiss the first few returns from Google because (1) Wikipedia is a for sure and no way can I reference that (2)Trash--sites that are nothing more than useless ads and (3)if it's a political search, I can definitely count on ideology tinged results.
If this search engine is as good as many hope it will be, I'm sure that Google will make Dr. Wolfram a financial offer he can't refuse and before we know it "Google" will have a brand new interface.
Although not a real google killer right now.
I see some areas where this kind of system have a real edge over Google:
- If you want medical information, for example on cancer, google has no way to qualify result set. This type of system will present already digested information to the user.
No move navigating over the pages to see witch one can help.
- technical information: I would love to be able to type:'How to configure my dlink x123 wireless router' and get the answer on the first try.
- a powerful understanding of the language used for the queries, will also lead to much better answers. Today we have to train ourself to think our searches in terms of keywords. Pick the wrong one and the results are not very useful; They seen to be able to understand a lot more than simple keywords.
Obviously time will tell if history remembers this particular product, their marketing team have done a good job in peaking everyone's interest, so at least it will get a fair trial rather than death by obscurity.
We should remember that semantic search engines of this sort are not new. While this one may be a significant advance beyond others, it is not ground breaking.
I'm not sure where this product sits in the field of search.
If I want an encyclopedic article, I would use a site like Wikipedia. It is peered reviewed unlike the results of a search engine and is written to present the information by humans for humans.
If I want to shop online, I will just type the product name directly into an indexing engine like Google. I don't need to type "I would like to buy X".
If I want to find the product manual for something I have bought, I would have to navigate the manufacturer's website myself, this information will not be reproduced locally by Wolfram.
The point is, what semantic search does is replace the "thinking" that I would normally do to contextualize my search. The thing is, I don't mind thinking, it's what I've always done and my brain is remarkably good at it.
Tim wrote "The thing is, I don't mind thinking, it's what I've always done and it."
You are in the minority.
"my brain is remarkably good at [thinking]"
Many people aren't so hot at thinking.
"I don't mind thinking"
Of those who can think, a fair number don't like to think.
"it's what I've always done"
Almost no one I know can say that (myself definitely included).
Not saying this thing will remake the world, but then again, look at some of the useless crap that remakes the world. Facebook is not a Nobel prize winning idea, but it's changing people's lives. So are Craigslist and Ebay. None of these things were the culmination of anything great, but nevertheless millions of man-hours and dollars are spent on them.
The unthinking masses matter. A lot.
I’ve just written a blog post about Wolfram Alpha in which I wonder about the possibilities this new platform could provide to open government advocates. Could you give us a clue into whether campaign and political data will be included in the computations Wolfram Alpha can make? Particularly, will it be able to make calculations on such things as political donations and roll call votes in Congress?
from what i saw, i can say google will remain on top.
To be honest, microsoft encarta (cd-rom version) is better, they just need a better web interface to that data.
This is like the data on the mathamatica website/products, combined with statistical information.
Great for scientists and scientific type research. A lot out there already exists in the arena which was originally sold on CD-ROM.
Is everybody just forgotten all the wonderful interfaces and the Multimedia CD-ROM revolution?
I can't feeling that is nothing new, and this data exists already on the web, and it could have been presented better.
I guess DELL paid all this hardware
The second video here looks like pure advertising, with cameras focusing more on the DELL logos than anything else...
WolframAlpha appears to be a new type of search engine, neither Wikipedia (based on social interaction of its experts) or Google. While useful, Google is basically an overglorified index that helps people find things on the web. Google "what is the weight of the Statue of Liberty" (comment #16 above) and you get 502K results). I look forward to testing it. Ted Downing, Arizona Research Laboratories, Univ. of Arizona
For those who think Wolfram Alpha is a threat to Google's supremacy in search technology, lets be clear about the fact that Wolfram Alpha works in a different sector. It is more of analysis and getting the usefull information instead of the website links. Of course it can be complimentary to Google. And in anyway, Google has been constantly improving it's search engine algorithms and way ahead of the competition.
Enjoyed the video especially.. This wolfram search is pretty cool :)
Wow - two whole server banks...I'm underwhelmed. Google probably has more computing power in their executive washroom.
this is obviously not a google or wiki killer because its not another search engine. it will answer any question that you want the answers for in the best possible way - a truly revolutionary idea (the idea that people who build computers had in mind).
it might be complimentary to a google search -:
keyword - 'holiday + india'
google = 1 million search results of which i might check only ten satisfying results
adwords = hotels, holiday packages, destinations etc
Alpha = everything about a holiday in india including availability of rooms in budget hotels, destinations according time of visit, which flights/trains/cabs to take, fares, rents, climate, ploitical scenario, forex, and everythng you can imagine.
so besides being a tutorial for any student it intends to solve the daily complexities of nature.
what we have to wait and watch is -:
how long can the keyquestion be and limitations of the computation that wolfram servers process. so the debate is not on the answer, which i believe will be the best, but on the key question.
It might answer a question for free but will it allow the commercial use of the answer - eg. it might help me put together a website but will i be able to publish it?
i might use a seriezs of questions and make a sizeable meaningful data of the answers but can i ...
and
will it be able to do things for me - eg. if i ask it to track an auction or a shareprice, will it be able to track and act on my behalf.(it would require it to track and then login with my credentials and take the action) if it does - whoa - we have a new world ahead of us...
i hope for the best and wish WRI all the best ...
will computational knowledge management software/algorithm(wolfsearch) help us find an answer to fight the crisis :)
i hope WRI is able to evolute into an industry like google and find answers both the commercial as well as the altruistic way ...
wondering what media solutions they will offer ... alphawords/ wolfwords/ werewolf or alphawolf
:)
All looks nice but as far as I can see it doesn't actually state where it's getting it's information from. So you are relying on the system choosing the best source. But who decides on the best source?
Also by just utilising the data from websites but not referencing it, aren't you just hyperlinking into sites and taking their data? I'm sure some countries have said that isn't allowed, especially as it's taking money away from sites by depriving advertisers of hits etc?
I know this will be a great system but I like to make my own decision of what sources I want to use.
Weather as an example. If I searched for the weather online I could get 5 different sites for example. I know that 1 of them has a sensor in the middle of my city, but another ones is over 30 miles away. The one in the city is a local university, but the one 30 miles away is a governmental website. Wolfram Alpha chooses the governmental website because it's 'official' say. But that information would be woefully inaccurate. I know that as I have looked at the sites to check this information.
I'm looking forwards to see how WA works but I'm at a bit of a loss as to how you could use it beyond basic encyclopedic or wikipedic information.
Definitely more like an encyclopedia than a directory of sites...thus competes with Wikipedia. Agree with others here in that Google or other search engines can approach this problem too.
Growth and success needs to consider localization as semantics in language is very nuanced.
It is ridiculous to even whisper wolfram alpha is a google killer, first of all, google is a company, with investements in many markets on the internet and outside of it, so this software can not compete with all google software, hardware and applications.
Wolfram alpha seems to be a mathematical software, of course it can answer many questions, searching for a website requires math too, so why am i going to use wolfram alpha instead of google search engine or anyother google product or service?
I do not think so, the ads that said google killer was only to spread thr word of wolfram alpha in a negative way in order to create impact on its product, does not work for me! I got to see it, evaluate it and judge it for myself.
To end this, wolfram alpha is new, so it will have bugs, we are talking of what it is but, personally i did not see wolfram alpha in action, in real action like i seen google products and services, besides, even if wolfram alpha is really really good, i would not use it, it is more like a software for advanced users or people that actually need all of that information, not for common people like me.
Besides all this, google search engine can provide the same information as wolfram alpha, not in one page of course and not in only one service, but i did not see wolfram alpha did it also, and i beleave they can not beat google in any way.
If you use google advanced operators you can achieve your search goal.
I have been following WA from March the 5th,and the more I learn about it, the more I am convinced that this is just another website for solving some well formulated Math question,and some science and economics questions. When it is launched hopefully this end of May, people who called it a Google killer will see how far they are from the truth, to compare an equation solver website to Google is a joke.
I really hope I am wrong and that we will find some way of using WA in a different way.
I'm sure everybody that uses the internet regularly have found ways to make it useful for themselves, e.g. favourite sites that we are interested in and already list everything we want, or ways to use google to find things in that "The more you pay the more clicks you gonna get thing"
I suppose I think the words "Ignorance is bliss" applies to users that see Google and the internet that way (probably us all me included).
I see the internet as a way to share information with the entire world instantly. Unfortunately having access to all the information in the world at the same time means that you have to search though all that billions and billions of gigabytes of information available, right.
My experience with google is that the answer I get when even doing a more directed search is probably representative of about a millionth if not less of all the infomation out there about the topic I was searching for combined.
So I think they have something going by giving a "Summary" type of answer. It is my understanding that this type of technology can also be used to direct you to the page that have the most relevance to your question. And what about possibly building a profile of how you are thinking, thus understanding what you meant by your question better.
Now google is now almost synonomous with the inetrenet, but remember google is not actually the internet, the interenet still exists without google, right.
So I think if WA can direct me to sites that have more relevance to what Ii am looking for, I will CAN google in a second, belive you me !
Maybee Wolfram Alpha is not perfect in terms of interface and the way the widest range of peolple would like to use it but I think they have something valuable. Theer is a considerable jump from crawling the web, making keyword indexes to searching the web making an index that has a deeper representation so I am by no way convinced that google will be able to do this, without WA's help no matter how much money they throw at it, maybe this is just enough to prompt more research into the field.
Time will tell
My point of view is, Wolfram Alpha amazing, spectacular, an revolution for Internet search's
is not like a Wikipedia or Google
I mean they both are differents. Google helps you find sites on the web, Alpha provides you with an "expert, detail, and especific opinion" the best answer for a good question.
especific questions, logical answers....dont you ?
Wolfram alpha has the potential to position itself on the web as a leader in this aspect of computing.
Wolfram may have ushered in a new era of computing - Engine computing.
I do not see this company as a competitive threat to google but a new era of engine computing should be of a concern to google as it has not demonstrated its lead in this aspect.
This technology can be revolutionary in a way that we have not yet imagine.Ibm's Watson is scheduled for launch around 2010 , The watson engine would face off with jeopardy in a live televise show with the aim of knocking off the human contestants in the jeopardy show.one could easily envision that ibm would put this engine on the web.
When technology like this emerge ,various players also emerge .We can look at what is happening in the cloud computing arena where we have Amazon.com, google , ibm,sun,etc.
hate your logic for publishing ..
It sounds good. Very complicated and obviously alot of thought went into this. It may even have a chance in my opinion.
it's a FIND engine!
This is going to be an exciting alternative to any other search tool on the web. It's just interesting that there are so many naysayers and chronic malcontents that seem to be dismissing this technology before it's even released for use.
I watched the entire demo presentation at Harvard and though I was a bit upset that it didn't show the screen or examples, I still think it's well worth watching and wrote up my thoughts on WolframAlpha and why it's NOT going to be the Google Killer: http://www.sagerock.com/blog/wolframalpha-not-google-killer/
Thank you for the screenshots and info!!
This is a step before the release of the 4 lines of code that will change the world as we know it.
Why not use blades? save a lot of space
This does actually compete with some Google functionality. For instance, you can type in "24 gallons to liters" and it will tell you the conversion. Or "Weather in Indiana" and it will give you a chart. But obviously, Google's implementation is much simpler than this. Google could buy WA and replace that functionality and be much better for it.
But there are questions about the neutrality of WA's data sources, and I don't necessarily like the fairly arrogant assumption that there is only one accurate set of data about the world.
Also, Wolfram Alpha's branding is shit, and will suffer because it's egotistical.
This is an intresting idea, the problem is that it doesnt appear to provide sources. How do we check the answers it gives? This provides an answer according to certain inputted data but we have no idea what data or whether its been entered accuretly
This is fantastic. I believe it is launching soon. Here's one with same shots as u - http://cafearjun.com/2009/05/15/wolfram-alpha-screenshots-and-my-test-reviews/
Today, I ran a test with these 10 questions across Wolfram Alpha, Google and True Knowledge.
(1.) Who discovered radium?
(2.) Where is Atlantis?
(3.) How do we make gold from lead?
(4.) Can robots dream?
(5.) What is a sprite? [This is my trick question since ‘Sprite’ is a drinks brand as well as a type of fairy.]
(6.) When did Homo Erectus become Homo Sapien?
(7.) Why are we here?
(8.) How many light bulbs are there in the world?
(9.) Who is the Vitruvian Man?
(10.) Where is Schrodinger’s cat?
These are all science-related questions which is where Mathematica (a reference algorithm of Wolfram Alpha) has been originally rooted. For the purposes of objectivity, it was only right to ask questions as close to a like-for-like basis as possible.
In several instances WA said it wasn't sure what to do with the input and didn't semantically connect the terms in the search query or generate relevant visuals. Google at least made some attempt to provide relevant and related links. True Knowledge, on not understanding the question, put the onus back on the individual to define the search term via its user wiki.
The result of Question 10 was particularly surprising. Natural expectation would be that WA would generate the integrals which form Erwin Schrodinger's postulations on quantum physics and compare/contrast these with Einstein's proof and Hawking's calculations.
It didn't generate any results and only produced a "Wolfram Alpha isn't sure what to do with your input."
There's definitely room for improvement and it's premature for it (and every other NLP/Semantic play) to position themselves as "Google killers" --- particularly given that Google Squared seems to be imminent.
All-in-all, though, competition is healthy for this space and great for us as information consumers, knowledge connectors and sense discoverers.
I am wondering something. If someone comes and types in an invalid website in the comment section, will it reduce the page rank of the blog it was posted on?
1 2 Next