A Look At the Wolfram Alpha "Search Engine" 216
An anonymous reader points out a ReadWriteWeb piece on an hour-long demo of Wolfram|Alpha (which we discussed at its announcement). Stephen Wolfram does not like to call it a "search engine," preferring instead the term "computational knowledge engine." It will open to the public in May. "The hype around Wolfram|Alpha, the next 'Google killer' from the makers of Mathematica, has been building over the last few weeks. Today, we were lucky enough to attend a one-hour web demo with Stephen Wolfram, and from what we've seen, it definitely looks like it can live up to the hype — though, because it is so different from traditional search engines, it will definitely not be a 'Google killer.' According to Stephen Wolfram, the goal of Alpha is to give everyone access to expert knowledge and the data that a specialist would be able to compute from this information."
search engine that supports pregex (Score:1, Interesting)
I'm still waiting on a decent search engine that supports perl regular expressions
Re:Google-killer? (Score:4, Interesting)
What happens when McDonalds is staffed by robots?
http://www.marshallbrain.com/manna1.htm [marshallbrain.com]
Great for financial data (Score:5, Interesting)
If they can figure out how to get this thing to understand financial data, it would be quite useful. That whole area needs more theoretical work.
Machine understanding of financial data is tough. Partly because the data is willfully obfuscated. I once developed a system for turning SEC filings into XBRL (which is an XML representation for financial statements.) At one point, I had several hundred euphemisms for "Net Loss". The connection between financial reporting and reality is at times tenuous.
Accounting is fundamentally mis-designed. The problem is that some numbers are actual, some have tolerances, some are estimates whose actual value will be known at a future date, and some are estimates whose actual value will never be known. Numbers of all four categories are added, and the result is given as a number without a tolerance. That's just wrong. Accounting works that way for historical reasons; it was designed when arithmetic was expensive. Why it stays that way is more interesting, but beyond the scope of this posting. Because of these problems, machine understanding of traditional accounting data is very difficult.
(Back when I did Downside [downside.com] I was more into this, but when I started getting invited to accounting conferences, I realized I didn't want to get into accounting standardization as a field.)
Re:My god, it's full of... (Score:4, Interesting)
Yeah [flickr.com], I'd say that's less than 10,000 CPUs.
That said, the later you try to crash the party, the more mature competition you are facing, and the bigger/better the launch has to be. Google didn't have Google to contend with.
Will it fail? Probably. But the stakes are enormous, so you can't blame a rich smart guy for trying.
Re:My god, it's full of... (Score:5, Interesting)
Google at the time had a.o. AltaVista to contend with, at the time the number-one search engine. It was set up by some college students in their dorm room, who had a better idea about searching/indexing web pages, and managed to implement that idea. Then it went live from a single computer for their friends. Who told their friends, and soon the whole campus used them, etc.
Google never advertised their service, it was pure word of mouth. They just got better results than the competition. And they got started of course in a geek environment, so the first word got out and spread quickly.
Good chance that the "next Google" starts up just like that. Hell, I bet The Pirate Bay started up that way. Craigslist did so at least - just a guy called Craig who started a local classifieds page for friends and friends of friends.
Yes the stakes are huge but just throwing money at the problem generally won't get you far, I would say good chance it gets you doomed even as big money often takes away the focus from the innovation that is needed.
This guy is indeed so full of himself (Score:3, Interesting)
Re:Great for financial data (Score:4, Interesting)
I think it is worse than that. They use those formulas because it prevents them having to think. I had accounting in a one semester stint in the Business School at a Large Unnamed University (before I realized that way madness lay). In the middle of an exam, I couldn't remember the Big Formula for solving one particular problem. So I derived my own, solved the problem with the correct answer. My formula was a special case of the Big Formula and my formula was just fine for the particular test problem. I got no credit, that's when I figured Business School Product deserved no credit.
Re:My god, it's full of... (Score:1, Interesting)
You could have said the same thing ten years ago. Alta Vista was run by DEC (Compaq), they had loads of servers, tons of disk space and millions of pages indexed.
Google didn't have any of that. But they had what mattered: Better search results.
At the time, Alta Vista would give around a million hits for any search term, and the one you were looking for was somewhere around page 50,000.
Google usually had the one you were looking for as the first hit.
They don't now. Too many people gaming the system, too many product comparison sites, too many forum posts with the same problem, but no answer, and so on.
A couple of people with a better algorithm would be able to beat Google, just like Alta Vista was beat before.
Not useful for the unwashed masses (Score:4, Interesting)
Not only is this not a Google killer (it's not even a search engine, so how can it be?!), I very much doubt it'll be of any use/interest to anyone outside of the intellectual elite. Googling for "swine flue" is of widespread interest, but I suspect that Alpha-ing for computed relationships and statistics is not.
http://www.google.com/trends/hottrends [google.com]
For anyone who has yet to read about Alpha, what it is is basically a large expert system written in Mathematica that computes the answers to queries covering a very large real-world knowledge domain. I havn't even read that it goes out to the web at all - it's basically based on a huge human collated/organized ("curated" in the Alpha parlance) data set of statistics and relationships. Apparently the results are presented in a very slick way including charts and graphics.
No doubt Alpha is a huge achievement in it's chosen domain of knowledge organization and computation, but I find it hard to imagine that a significant portion of the population will find it useful.
Re:Brother . . . (Score:1, Interesting)