| 8:29 pm on Mar 10, 2009 (gmt 0)|
I tried reading Wolfram's 'A New Kind of Science', and it was way, way, way over my head. What I did gather from his book, though, is that most of what we see as complexity in nature is actually the easy-to-compute by-product of a relatively small number of algorithms at work.
I'm no math-whiz, but given that DNA itself is composed of an exceedingly small number of building blocks, it makes sense to me that we can hope to find the simple math underlying complex systems and reverse engineer our way to complex but correct answers to complex questions.
| 10:05 am on Mar 12, 2009 (gmt 0)|
Not much information there but it sounds as though they're planning to answer the questions on their site rather than sending people to sites that they've crawled.
If that is the case then I cannot see it working. They won't be worried by the few technically knowledgeable people who block them for consuming resources with no benefit returned. The ones that selectively feed them incorrect information will be a serious issue though, especially in niche areas without too many sites dedicated to them.
| 11:36 am on Mar 12, 2009 (gmt 0)|
The question is how to know exactly what people is looking for. People can use keyword or natural languaje but computers always will feed them with that they have told, not what they think are telling.
I.e.: the today's search with keywords "dogs" will give the same result that a natual language search like "I want to know about dogs". Only when the searcher specifies "buy dogs food" or "I want to buy food for my dogs" both search engines can give good results. I think the natural speaking way is still a hard way to go and maybe only will be usable when computers are something they aren't today.
A good friend who introduced me in this computer's world always say: "Computers do exactly what you tell them to do, not what you want them to do".
| 7:07 am on Mar 13, 2009 (gmt 0)|
Wolfram's celluar automata approach to computing has a powerful and elegant depth that just might stand everything you think you know about computing on its head.
I've been recommending "A New Kind Of Science" to people for years - no, it's not the easiest read (and it's VERY long, too) but it really does introduce a paradigm that could change everything about how we do science and computation. There are some fundamental assumptions at the root of science that may just be untrue, and Wolfram sees the way around those potholes.
I look forward to what this approach may offer. Wolfram brings fundamental genius on the level of Einstein and Hawking, rather than intelligence at the level of Page and Brin.
| 3:57 am on Apr 30, 2009 (gmt 0)|
For those who can take in a nearly two hour video - here's the sneak preview webscast of Wolfram¦Alpha from Harvard's Berkman Center.
Warning - this is a poorly shot video from an visual information standpoint. Wolfram is clearly demonstrating things on a screen in the room that is never shown.
| 8:30 pm on Apr 30, 2009 (gmt 0)|
System: The following 2 messages were spliced on to this thread from: http://www.webmasterworld.com/alternative_search_engines/3904367.htm [webmasterworld.com] by bakedjake - 9:55 am on May 1, 2009 (edt -4)
Wolfram Alpha Computational "Knowledge" Search [news.bbc.co.uk]
|The "computational knowledge engine", as the technology is known, will be available to the public from the middle of May this year. |
"Our goal is to make expert knowledge accessible to anyone, anywhere, anytime," said Dr Wolfram at the demonstration at Harvard University's Berkman Center for Internet and Society.
The tool computes many of the answers "on the fly" by grabbing raw data from public and licensed databases, along with live feeds such as share prices and weather information. Other functions solve complex mathematical equations, plot scientific figures or chart natural events.
| 8:53 pm on Apr 30, 2009 (gmt 0)|
I've been watching this pretty closely. I can't wait to see what they come up with.
Here's something that I found very interesting, their team assembled a timeline of computable knowledge:
| 10:11 am on May 2, 2009 (gmt 0)|
Anyone remember that story by Isaac Asimov.. the one about a computer called Multivac? Remarkable similarity I thought.
| 11:37 am on May 2, 2009 (gmt 0)|
It might turn up very useful for science students and wider scientific community but this approach will not replace full text keyword driven WWW scale search engines.
| 4:19 pm on May 2, 2009 (gmt 0)|
|I'm no math-whiz, but given that DNA itself is composed of an exceedingly small number of building blocks, it makes sense to me that we can hope to find the simple math underlying complex systems and reverse engineer our way to complex but correct answers to complex questions. |
This is not how the real world works. Lets give a simple example. It is safe to say that Einsteins theories can all be written down with the first 127 characters from the ASCII set.
Yet you won't be possible to deduce Einstein theories once you know all 127 characters, not even if you have a library of 1000ths of books where all these 127 characters are used repeatedly. Characters are just a method of representation but give no information about the language (syntax) nor the meaning (semantics) of Einstein theories.
The same for the Wolfram Alpha project. It may be capable of better parsing and understanding existing knowledge and present it to the user compared to flat search engines, but creating new knowledge is something completely different.
| 4:45 pm on May 2, 2009 (gmt 0)|
|with trillions of pieces of curated data and millions of lines of algorithms |
Does WolframAlpha include a transient mood algorithm with binding to facial recognition algorithms ("Do you really hate me or are you just having a bad day?") and an unspoken assumptions algorithm?
That would be helpful. :P
| 3:54 am on May 3, 2009 (gmt 0)|
"All one needs to be able to do is to take questions people ask in natural language, and represent them in a precise form that fits into the computations one can do."
While with the Mathematica project, he and his fellows contributed great deal to new algebraic innovations and indeed understandings, Al-Khwarizmi himself would be proud of, so would Einstein. A lot of what they have achieved is now implemented and applied in many fields. However, the NKS and the Automata argument does open a healthy debate which will push for further achievements. Computing human emotion, judgment and feeling is what needs to be researched and logically understood more. The day when two computers talking to each other in natural language and one asks the other: How are things? from that question the other computer would be capable of understanding that question meant: how are you? or how is life treating you?...and if the answer was: Up and Down, the questioner also needs to understand that it meant some times I am OK and sometimes not, having in mind all the profile data of that machine's "personality".
Natural language understanding is not enough, it is easy to have computers trained to actually talk to each other in plain language, compute and assess each other's questions in regard to preparing an adequate answer based only on factual computational linguistic algorithms which are themselves based on external profiling of that (ad-hoc) entity (ranking, importance, age, gender, quantity, trust, location, wealth, political affiliation, habits....), still used, added to and improved by Ask Jeeves and Google. The Alpha project is slightly different, but nonetheless looks to be similar, and is yet another algebraic solution. The problem is, that answer should be based on billions of readily available emotional triggers. One to One conversation is not actually two entities ONLY talking to each other, it's a Soul To Soul interaction and that's what should be worked. The latter involves computing the biological, psychological, physical and physiological "state of the mind" of a given entity before any answer was considered, natural language then is only a tool to communicate that answer.
I am not knocking the Wolfram Alpha project, what it claims to offer in terms of "Common Sense" answers, ask jeeves (known as ask now) are pioneers in this field, google does pursue that model right now, so did alta vista and a host of other search engines over the years to a lesser success. From all the "Common Sense" answers Search Engines, while far from perfect, probably the Ask module is the most reliable, you can test it. Whether the Alpha project will be able to surpass Ask's genius, that remains to be seen.
If Ask Jeeves concentrated more on clever promotion, Google would not have had a chance for dominance, they instead concentrated on the core product instead of marketing themselves properly and their idea. The Alpha project will in my view suffer the same consequences as Ask Jeeves's.
| 6:19 am on May 3, 2009 (gmt 0)|
IMO this will be the search version of Wikipedia, not the "find lyrics of ...." type. As as I can tell, they appear to work from databases, not by scanning the net.
Google does this thing a bit: stock prices, forex, movie tickets, from info CIA country list etc. Now they just have to mix them
| 1:49 pm on May 3, 2009 (gmt 0)|
It will be exciting to test this when it becomes available. But it does sound more like a fact engine than a search engine, and that fundamental difference cuts out the opinion-based web and ecommerce.
If you rely on getting traffic by providing facts, this may be just the next nail in the coffin. Search engines have been moving in on this territory for ages with attempts to answer questions at the top of certain serps.
| 11:14 pm on May 3, 2009 (gmt 0)|
Imagine a person who has seen everything, done everything and has retained all possible knowledge from each and every single event.
That's Google - able to retrieve whatever you're looking for from an enormous database as quickly as if you were asking this person face to face.
Wolfram seems to be a different beast, not only is all the knowledge there but it's being analyzed and compared in some ingenious ways.
The only flaw that I can speculate on is that the end user, the human performing a search, will just not have and perhaps never have the ability to grasp whats being done.
The video is a must watch, poor angles aside, but until society values Wolfram (28,000 views) more than Britney Spears (a bazillion views) I don't think the ambition behind Wolfram can truly make enough of an impact to make significant change. Thanks for the video link.
| 11:26 pm on May 3, 2009 (gmt 0)|
From the BBC story:
|For example, a relatively simple search, such as "who was the president of Brazil in 1923?", will return the answer "Artur da Silva Bernardes". |
So does G, and all-comers, so where exactly is the sausage in this bit of media sizzle?
| 5:29 am on May 4, 2009 (gmt 0)|
The video that tedster posted is really worth watching. Initially I thought oh that's really cool, but who would use it and what would they compute? But as you watch the video with your rampant capitalist hat on that we all have the list becomes endless.
Once people learn how to use it, I think it will prove a real shift in productivity in a lot of industries.
| 11:32 pm on May 5, 2009 (gmt 0)|
Mathematica always amazed me. I am sure this one has the potential to do the same. However, it will probably be another intelligent publisher not a search engine.
| 9:25 am on May 7, 2009 (gmt 0)|
Wolfram's really should do a talk on the Ted stage - I'm sure he will get lot more appreciation of the technology. He would have create a new programming language just make this language conversion translate efficiently to logic.
I'm wondering how this is going to perform in the real world. As the search engine (well, not really a search engine), seems be very clever only on certain subjects.
| 10:59 am on May 7, 2009 (gmt 0)|
I think there's a lot of appeal of this approach for a significant amount of searches. As an example, if I need to know the current time somewhere, I search [current time [country name]] into Google, and get a precise answer back - I don't need or want any lengthier response than that.
| 12:05 pm on May 11, 2009 (gmt 0)|
More today on CNN [scitech.blogs.cnn.com]
[edited by: eelixduppy at 7:36 pm (utc) on May 12, 2009]
[edit reason] fixed URL [/edit]
| 4:11 pm on May 15, 2009 (gmt 0)|
Launch scheduled for today, you can watch the launch online. More than 10 000 CPUs, 5 datacenters and a huge software infrastructure. That is going to be very interesting ! Starting at 8pm EST and final release for monday. To watch closely !
[edited by: Cirilo007 at 4:12 pm (utc) on May 15, 2009]
| 11:47 pm on May 15, 2009 (gmt 0)|
To see the broadcasting :
| 10:48 am on May 16, 2009 (gmt 0)|
It certainly is a promising concept. Nice mashup of data. Could be the death of quite a number of informational sites.
It certainly isn't perfect yet. A query for Washington returned a city in the UK. I expected either the president or the larger city in the US.
| 12:40 pm on May 16, 2009 (gmt 0)|
mm.. not very backward browser friendly at all.
And currently experiencing over traffic issues.
Has some issues to ion out:
[edited by: Seb7 at 1:13 pm (utc) on May 16, 2009]
| 1:07 pm on May 16, 2009 (gmt 0)|
I've been checking it out for some time, seems like a really promising concept.
Today I finally tried it and it almost crashed my firefox, lol : D
"A script on this page may be busy, or it may have stopped responding. You can stop the script now, or you can continue to see if the script will complete.
Lol, Jquery : D
| 7:28 am on May 18, 2009 (gmt 0)|
I am actually sick of this "new google" kind of thing everyone is hoping for. Answer questions, isnt answers and wikipedia enough? This is huge waste of time.
I think Google is king and has completely evolved ridiculously more then a "new concept in search". The problem is we do not breakup the analysis of changes since there inception. They have so many new ideas they have integrated in search behind the scenes including webmaster central, analytics, maps, blogs etc just to name a few. They have the data to actually know which direction search is going on. They not only have to find 20% new pages with every crawl, they have to combat spam, algo flaws continuously. To say we answer questions sounds like a concept from 10 years.
How about this for a new concept, how about actually try to not be an answers.com, wikipedia.org and try to be different. I dont see this as anything but as a waste of time. I will leave it to google to organise the worlds information.
| 7:52 am on May 18, 2009 (gmt 0)|
At least you can't say, mathematicians do not have a sense of humour:
|I am actually sick of this "new google" kind of thing everyone is hoping for. Answer questions, isnt answers and wikipedia enough? This is huge waste of time. |
Perhaps you should have invested 30 seconds of your time to check out wolframalpha before starting your rant. You would then have noticed that this is a completely different thing than google or wikipedia.
| 8:58 am on May 18, 2009 (gmt 0)|
..and this was discussed on NPR (with Danny S.). Link to audio:
| This 58 message thread spans 2 pages: 58 (  2 ) > > |