On that note, any other HNers planning to be at the demo in Cambridge on Tuesday? http://cyber.law.harvard.edu/events/2009/04/wolfram

It seems like it's worth going to - and it's directly related to something I'm doing at work (inference over topic maps).

I'd like to, but it's full! I'll be watching the webcast ...

Wolfram Alpha doesn't compete with Google, it competes with Wikipedia.

I expect students who use and quote Alpha as their source of information won't get the same scorn than the ones who quote Wikipedia.

It will be an interesting battle: crowdsourcing vs. proof.

And perhaps Freebase: http://www.freebase.com/

I think they supplement each other. Wikipedia is a great place to go explanatory material; statistics are a part of that but if you are really interested in the numbers behind something you'll probably follow up the citations. Alpha looks to have great potential as a tool for statistical discovery and analysis, but I rather doubt it auto-generate tutorials on what data mean or why they matter.

It would be nice if Wikipedia can improve the quality of its articles by linking to Alpha; I hope Wolfram research won't exert a propriety attitude to its search results, but will make them available with something like a creative commons license to the extent that the original sources are in the public domain.

Quoting Wikipedia is not respectable in an academic context because you are not quoting a primary source, or even a secondary source. You are typically quoting a tertiary or quaternary source. If the article is properly cited, however, it is not difficult to simply quote the original citation.

this sounds like something that spans the (rather wide) gap between google and aws + public data sets. I wonder how far you can go in terms of combining data/hacking on it?

For example they mention real-time financial data. So can I write something that would connect online discussion/activity to stock movements?

A platform for doing large scale, real-time data analysis in a rich descriptive language without doing any of the dirty work would be quite something.

This is really interesting on many fronts... Google queries are trivial and dumb - they take virtually no CPU power to execute (on a per-query basis) in the grand scheme of things. But I'm trying to imagine how Wolfram Alpha will scale, and I'm not really seeing it. You can't precache results, queries require the aggregation and manipulation of huge data streams, and so on and so forth.

Wolfram Alpha is going to be _expensive_ to keep up. It's good to see they've already thought of this (with their Pro offerings), but the thought of the sort of power this thing would consume if it becomes popular en masse makes me shudder.

Expect bugs and funny search/result pairs to show up in blogs soon. But this is a great contribution to humanity.

Didn't knew that something like that could be coded in Mathematica.

Indeed, Wolfram himself once solved a Google puzzle (to encourage job applicants) using Mathematica. It was, what's the pattern here, and what's the next row?

1 11 21 1211 111221 312211

I always find these puzzles odd

Why can't this row be valid?

1 11 21 1211 111221 312211 1 11 21 1211 111221 312211 1 11 21 1211 111221 312211 ....

Why does it have to be 13112221? Is it just a most close answer according to average math pupil's knowledge model psychologically?

Yeah, the "next # in the sequence" things are in theory dubious but in practice the theoretical unsoundness doesn't really cause issues (in the sense that: the people you'd want to select for will often find the pattern you want them to find, and many of the people you want to select against won't find that pattern).

The real meaning of what they're asking for is something like:

- I assert that there's a simple-but-nontrivial "algorithm" that generates this sequence. Find that algorithm, and use it to find the "next number in the sequence".

...but then you need a definition of "trivial", which is a lot of work (and most applicants know "trivial" when they see it:

  def sequence(i): return [1,11,21,1211,111221, 312211][i % 5] 

...is "trivial", so why bother explaining it that way when most applicants already "know" that's the wrong answer?

Doing a more-formal definition probably gets you into issues of Komolgorov complexity, which most applicants will not know much about.

Google is smart enough for this:

http://www.google.com/search?q=1%2011%2021%201211%20111221%2...

But why Google when you can OEIS? http://www.research.att.com/~njas/sequences/

This is an awesome resource.

Only after they posted the puzzle did solutions pop up on their own search engine. :)

WOW, this made my day... this is AWESOME. Can't wait to try it out!!! I'll def be watching the webcast.

I really hope this works good!

The first question I'll ask WA is:

"What is the answer to life, the universe and everything"

Don't disappoint me...

I'd ask it for the perfect recipe for a decent mug of tea. It'll hold up WA's severs for months on end, causing unmaskable interrupt that won't return until I get my tea.

"Can entropy ever be reversed?"

I'm actually going to be really disappointed if Alpha doesn't have the right answer to this.

I'm thinking of a number...