Author Topic: Wolfram Alpha - Impressions On Launch (Read 41,363 times)

raybeere · « **on:** May 16, 2009, 04:05 AM »

It is late, and I'm tired, with a nasty developing headache, but I'm sure some DC readers are aware of Wolfram Alpha and are curious. I had the page loaded, with the nonworking "Launching in May" announcement, and when the announced they'd activated it, I refreshed and got in as soon as the page reloaded. It took a few seconds. Rather than rewriting all my thoughts, I'll link to what I posted on my blog, if that's okay. The post is here.

Yes, I was pretty harsh on them, but my perspective was that of a user / researcher. This is supposed to be something new, something special, and except for math problems (and how many days, weeks, years, etc. since a particular date is just a specialised math problem) it doesn't seem ready to do anything at all. If anyone is really curious, I do have screen shots of the searches I mention, although not the ones I just summarise. (I tried "What happened in 1666?", for example, and "What happened in London in 1666?")

mouser · « **Reply #1 on:** May 16, 2009, 04:30 AM »

I have a long standing dislike of wolfram after reading "A New Kind of Science" and questioning him at a talk he gave.. Another self-deluded egomaniac if you ask me.

Now that's not to say that "Wolfram Alpha" is not going to be a great thing -- only that i have little respect for him.

superboyac · « **Reply #2 on:** May 16, 2009, 05:14 AM »

Quote from: mouser on May 16, 2009, 04:30 AM

I have a long standing dislike of wolfram after reading "A New Kind of Science" and questioning him at a talk he gave..

Care to give details? I also sensed that somewhat from what little I know about him. PM me if you don't want to be public.

rgdot · « **Reply #3 on:** May 16, 2009, 06:09 AM »

Atomic weight of Chlorine...worked
Biggest town in Germany...failed

Just two random tries that, on google, either work directly (answer on top of search results) or provide a good link with the answer.

cyberdiva · « **Reply #4 on:** May 16, 2009, 01:39 PM »

Quote from: raybeere on May 16, 2009, 04:05 AM

Rather than rewriting all my thoughts, I'll link to what I posted on my blog, if that's okay. The post is here.

FWIW, I thought your blog entry was terrific! It was interesting, informative, and--in spite of what you claimed--well written. It also prompted me to try Wolfram Alpha. I spent much less time than you did, but Wolfram Alpha struck out on all my queries. They were simple queries--"who is _______?" for example. I tried my husband's name, my name, and my brother-in-law's name. Google had no problem returning information on all of us, but W.A. didn't know what to do with any of the queries.

I tried several more queries of a different sort, and ended with "What is Donation Coder?" All I can say for W.A. is that it's consistent--it consistently didn't know what to do. By contrast, when I then put "what is Donation Coder?" into Google, its first three hits were to the web site, the forums, and Wikipedia's coverage of DC. Guess I'll stick with Google.

mouser · « **Reply #5 on:** May 16, 2009, 01:49 PM »

raybeere, i've been browsing through your blog entries after reading cyberdiva's comments -- and am enjoying your writing as well.

czb · « **Reply #6 on:** May 16, 2009, 02:05 PM »

I like it

I think it has got a great potential!

Maybe it is a first step of full Mathematica port to internet?

mouser · « **Reply #7 on:** May 16, 2009, 02:15 PM »

Quote from: superboyac on May 16, 2009, 05:14 AM

Quote from: mouser on May 16, 2009, 04:30 AM
I have a long standing dislike of wolfram after reading "A New Kind of Science" and questioning him at a talk he gave..
Care to give details? I also sensed that somewhat from what little I know about him. PM me if you don't want to be public.

well.. it's been several years since i read the book, so my memories are a little bit hazy, but the central premise of the book might have been controversial 100 years ago, but wolfram spends 100 pages building slowly up to an idea that he is sure is so revolutionary that the reader won't possibly be able to grasp it: Complex behavior can emerge from the interaction of simple rules. There's nothing controversial or surprising about that claim -- it's a completely accepted idea.

Wolfram blatantly refuses to acknowledge and site huge bodies of work on dynamical system theory and chaos theory, probably because doing so would illuminate how old these ideas are that he is trying to claim are the foundation of his newly invented Wolfram Science. This is what happens when you self publish, don't use an editor, and don't feel obligated to be responsible to the existing literature.

On a more subtle note -- I think there is an area where he is fundamentally wrong about, which i talked to him about. He focuses on, and is most interested in, interactive rules that produce unpredictable behaviors. He rehashes the emergence of some well-known dynamics that emerge in the natural world, like the beautiful fractal patterns found on shells -- and wants to extend such unpredictable complex behaviors to other areas of physics and the natural world. I think a fundamental and subtle problem with this is that on a macro-level, the real world is HIGHLY predictable and regular -- and that fact is essential to our successful and intelligent interaction with it. Since I come from an AI/Machine Learning background, this is a particular point of interest for me. The very kind of extreme pseudo-random emergent behaviors that he is so excited about and thinks are the most important, are the kinds i think are least interesting and least relevant to understanding the natural world. There are beautiful mathematical issues that came out of the work on chaos theory, like attractor dyanmics, that *are* significant and important -- and are completely ignored by Wolfram. Very very disapointing book.

Coeluh · « **Reply #8 on:** May 16, 2009, 02:49 PM »

It is awesome!

mouser · « **Reply #9 on:** May 16, 2009, 02:51 PM »

Shouldn't we have a link to Wolfram Alpha in one of these posts??

Here it is:

http://www.wolframalpha.com/

I tried it -- i think it has potential and it's pretty fun to use.

mouser · « **Reply #10 on:** May 16, 2009, 02:54 PM »

An an aside: How about someone make us a farr alias to search wolfram alpha?

Eóin · « **Reply #11 on:** May 16, 2009, 02:54 PM »

Mouser, not wanting to really get into a debate, I'm not particularly convinced by his work either. But the 'new kind of science' is about applying cellular automata to the very microscopic; to quantum theory and the most fundamental of the currently know particles.

Down there the world is exclusively random! The predictability of the macroscopic world stems from the idea that the interactions compound the randomness so much that you end up with chances that the 'weirdness' being seen on our scale drops so low as to makes such things unlikely to occur even but once in the lifetime of the entire universe. Though experiments have been devised to make such events observable, and they've succeeded.

My understanding is not that emergent behaviour is new, but that Wolframs research is it's attempt to provide an alternative theory of that world to say String Theory and the other lesser know theories.

mouser · « **Reply #12 on:** May 16, 2009, 03:16 PM »

Ok now you're making me get down my copy of the book from the top shelf

Quote

But the 'new kind of science' is about applying cellular automata to the very microscopic; to quantum theory and the most fundamental of the currently know particles.

I don't think that's a fair statement about the book. There is one chapter on physics near the end of the book -- i don't view it as the focus of his work. Go re-read chapter 12 where he wraps up the book and talks about his grand principle. I think the physics chapter was a throw-away thought experiment.

In fact i am completely unqualified to speak about his speculations on cellular automata physics. On a gut level, it feels to be like total speculation and hand waving which is likely to be considered trivial by hardcore quantum physics/stringtheory people -- but on the other hand, the basic core bare concept of the very lowest level physics being a CA has some intuitive appeal to me (but again i think his speculations beyond this basic idea seem unfounded), and is a non-trivial assertion worth exploring.

Quote

Down there the world is exclusively random!

Not to get into a prolonged discussion of an area that i have little expertise -- but i think the point is important enough to make at least one more pass. The issue isn't whether randomness is a fundamental low-level property -- indeed i find the idea of a deterministic CA generating what appears to be quantum randomness infinitely more appealing at a gut level than the "true" randomness that quantum theory suggests -- and surely so would einstein. [One of the big holes in my understanding of quantum physics is that i don't understand why quantum physicists are so convinced that the "apparent" randomness isn't the result of a pseudo-random deterministic low level process].

But regardless -- the point i was trying to make might be better explained by differentiating between what i think Wolfram is saying, and what i would say:

Wolfram: The interesting processes in the world are the result of the emergent interaction of simple rules -- and the key kinds of rules to look at are the super simple ones that lead to essentially-unpredictable super-complex output behaviors.
Me: The complex behavior we see in the world is generated by the emergent interaction of a very special kind of simple rules -- ones where the macro level behavior is extremely predictable and non-fragile, but also manipulateable (i.e. with important attractor dynamics, saddle-points, etc.).

mouser · « **Reply #13 on:** May 16, 2009, 04:25 PM »

Video of Wolfram himself discussing goals for Wolfram Alpha search engine:
http://www.youtube.com/watch?v=5TIOH80Qg7Q

Eóin · « **Reply #14 on:** May 16, 2009, 05:15 PM »

I guess I'm somewhat comfortable with the idea that it may all be random, but we do not have an explanation of the level below what we've observed thus far to know for sure it's truly random. It will be interesting to see if these things get explained in our lifetimes. Otherwise I'm inclined to agree with you that his new kind of science isn't all that new.

Anyhoo, not the most on-topic of discussion

mouser · « **Reply #15 on:** May 17, 2009, 04:00 PM »

just saw this:

http://negatendo.net/blog/2009/05/16/wolframalpha-and-screen-scraping/

Quote

Perhaps the oddest thing about Wolfram|Alpha is that the text that appears in query results is not text at all, but is in fact made up of dynamically generated GIFs

I didn't notice this the first time but sure enough.. you can't select and copy text in the results!!

The article goes on:

Quote

Perhaps the oddest thing about Wolfram|Alpha is that the text that appears in query results is not text at all, but is in fact made up of dynamically generated GIFs:

The Wolfram|Alpha FAQ claims:

Quote
All output content is rendered as images, for consistency.

Of course, a sentiment like that would make any web designer jump off a bridge. Considering all the other nice UI nuances that Wolfram|Alpha has, I call bullshit. It’s not about visual consistency.

I think it’s an attempt at preventing what is quickly becoming the bane of any informative website’s existence: screen scraping.

Screen scraping is, of course, using a script or bot to extract data from the visual output of a page. A web screen scraper digs out the needed data from the HTML source and formats it accordingly. This technique is used to subvert APIs, feeds, etc. when these “legally provided” methods of access don’t give you what you need. Of course, it’s also a way to really piss the people who run the website off, because screen scraping is typically immune to throttling and data control, unlike an API or feed, which can be monitored and cached.

And notably, it appears that this kind of control is of the utmost interest to Wolfram|Alpha, as it’s part of their “Step 3: Profit!” plan (also from the FAQ):

Quote
Subscriptions will be available in the near future with enhanced features for large-scale and commercial use.

What ridiculousness.

Grorgy · « **Reply #16 on:** May 17, 2009, 08:24 PM »

While i was seriously underwhelmed by this thing, I just found this little article at mashable.com ( http://mashable.com/2009/05/17/wolfram-easter-eggs/ ) . Some of these are rather good, ask it for example, the meaning of life and it will answer 42

app103 · « **Reply #17 on:** May 18, 2009, 12:45 AM »

Quote from: cyberdiva on May 16, 2009, 01:39 PM

I tried my husband's name, my name, and my brother-in-law's name. Google had no problem returning information on all of us, but W.A. didn't know what to do with any of the queries.

I tried that too, and it suggested to use only my last name, which returned statistical information from US census reports:

Quote

I tried several more queries of a different sort, and ended with "What is Donation Coder?" All I can say for W.A. is that it's consistent--it consistently didn't know what to do. By contrast, when I then put "what is Donation Coder?" into Google, its first three hits were to the web site, the forums, and Wikipedia's coverage of DC. Guess I'll stick with Google.

Instead of asking the same question that you did, I just typed in donationcoder.com and did get results.

From what I understand, you can't enter in any query (beyond mathematical) that it doesn't already know where to get the info from. The less you ask, the more you get. Single words return the most results, although most of what it returns is more or less trivia and not really useful.

It's more of an entertaining toy than a useful tool.

Quote from: czb on May 16, 2009, 02:05 PM

I like it I think it has got a great potential! Maybe it is a first step of full Mathematica port to internet?

webMathmatica is what it's using.

Quote from: Grorgy on May 17, 2009, 08:24 PM

While i was seriously underwhelmed by this thing, I just found this little article at mashable.com ( http://mashable.com/2009/05/17/wolfram-easter-eggs/ ) . Some of these are rather good, ask it for example, the meaning of life and it will answer 42

Take a look at some more of the searches people have done with it and the various impressions of users, from Friendfeed. (which incidentally, makes it easy to share the image based results, with their little bookmarklet)

Ehtyar · « **Reply #18 on:** May 18, 2009, 02:06 AM »

An excellent idea, and a very impressive implementation. If you think of it for what it is, I think you'll appreciate it, but all this "Google killer" talk is absolute nonsense.

Ehtyar.

[edit]
"Wikipedia killer" might be closer to the truth, but Wolfram is really too unique to think of it in these terms IMO.
[/edit]

mahesh2k · « **Reply #19 on:** May 18, 2009, 03:01 PM »

I asked few questions and i got weird answers and sometimes no results returned.

-What is the speed of light?
-Where is gallactus?
-What is the purpose of life?
-How to cook Mexican pizza?
-What is the distance between moon and earth?

Wolfram alpha needs to complete graduation before thinking about killing Google.

fenixproductions · « **Reply #20 on:** May 18, 2009, 03:39 PM »

I've tried Wolfram and it is sad experience. Many countries does not even exist in it.

I must say that one of my colleagues wrote Polish language chat bot for his diploma (took him year or something) and I remember it working better than Wolfram. His project was giving correct answers for questions like "When can I go from Warsaw to Krakow". Learnt database had only the size of few GB but in case of no data it simply asked Wikipedia or Google. It worked great that way. Unfortunately, implementing Polish semantic and grammar mechanisms was too time consuming work and because of that there was no English version

I can't simply believe that something such "big" can work so bad. Wolfram reminds me Chandler somehow.

40hz · « **Reply #21 on:** May 18, 2009, 03:40 PM »

I gave it a try.

1) My impressions after using it: Interesting. With a lot of work, it might actually be useful...someday.

2) My take on most of the claims currently being made for it:

[attach]

herneith · « **Reply #22 on:** May 18, 2009, 04:52 PM »

How the Mighty have Fallen!

gpetrant · « **Reply #23 on:** May 18, 2009, 07:59 PM »

Thought I'd add this as a supplemental to the discussion: http://www.pcmag.com/article2/0,2817,2347236,00.asp (I have no interest in the site at this point, beyond its super cool name (a biased opinion for obvious reasons

))

(My stance on new things in general: Darwinian. With very few exceptions, I am content to wait and see how well an app or service 'survives' (read: works for other users) first, before picking it up for a try-out. If it outlasts its initial (and hardest) months/years, that's a big plus in my books.)

gpetrant · « **Reply #24 on:** May 19, 2009, 02:47 AM »

Oh god! oh god! stay away from Wolfram Alpha..|.for god's sakes.,/ I...I can't explain everything right now.. it's all so...so surreal. I...I decided to do some research on Wolfram using Methematica (Crystal) and andsomehow my WiFi picked up *their* satellite feed and...and...oh god...I gotta go I hear something outside my window...please pray fo.........................................*syslog: transmission and transmitter terminated 05:10 hrs. Press Enter to reboot now.*

Hello, my name is Edna. I love Wolfram Alpha. I use it everyday for all my shopping needs. Yes, I love Wolfram Alpha. (And you'll learn to love it too!)