Hacker News new | past | comments | ask | show | jobs | submit login
Wolfram Alpha to Launch Live on Justin.TV (technologyreview.com)
50 points by abstractbill on May 13, 2009 | hide | past | favorite | 42 comments



In the meantime, they just released a screencast demo done by Stephen Wolfram. It seems to be getting lost in the other news of the day but its pretty cool.

http://blog.wolframalpha.com/2009/05/13/stephen-wolframs-int...


Whoah! I've used Mathematica a lot and I had high expectations for Wolfram Alpha, but this... was just incredible. Thank you so much for posting it.

I wonder if they'd consider opening the system up for others to develop 'plug-ins' covering more knowledge areas.


After seeing that, I strongly suspect my browser will get a 3rd text box next go the google search.


So... at about 3:30, he enters "6000 words". One of the calculations is wrong (screenshot: http://byrneseyeview.tumblr.com/post/107322432/i-was-just-wa...).

Am I missing something? Did the Mathematica inventor do a screencast with a multiplication error? Or is he using a video / mockup rather than an actual Alpha?


There are 5999 spaces between words. But that still leaves him 1 character short.


Unless, of course, he's assuming a 6000-word sentence with no punctuation except the period at the end.


That is some really amazing stuff. I can't wait to play with it myself.


one thing that strikes me how can it logically say "the gdp of france / italy" should be interpreted as "the gdp of france divided by the gdp of Italy" that makes no sense to me.... clearly the context suggests that I want to know the GDP of both countries side by side.

Actually im underwhelmed by it :( which is sad.


because it interprets "/" as mathematical division and "vs." as comparison.

they chose to use some specific, static operators so that you can do interesting things with the data, instead of having it simplify your request to what you "probably" want.


how many non-mathematicl non techy people (who realistically might want data like this, say, for a school report) would use vs. rather than /. it's just not natural english.

I understand your point: and I would have figured it out. But this is marketed as an everydfay search engine - which it's not, it's an advanced search for an encyclopedia :)


how many non-tech people would even assume that the search engine could produce this type of comparative results without learning about it first? in that process of learning, you learn how it works, and learn that /=division, vs=comparison

i imagine if it turns out to be a common problem, they'll probably give you a little reminder message about it. "if you're looking for data comparison, use vs. instead of /"


That little learning for technically literate (small percentage of users) could be a show stopper for the rest. Many people are scared of even the simple maths (sad but true).


Isn't it silly to judge a huge system by a single query which happened to to one logical thing instead of another?


That is an interesting interpretation, though I think its correct to assume that you are trying to specify a computation rather than presentation.

In general its best to thing of that input field as a API with a huge syntax space, rather than something that is actually "understanding" anything. Getting the side-by-side comparisons can be done in many ways, like just using a comma instead of /


my main point is that it is the "understanding" idea that WA is being presented as (either by Wolfram or the media I dont know). It obviously isnt: so people will get confused...


He mentions that simply putting items in a series would generate comparisons, so I think you'd get what you want by taking out the / operator.

(edited, no need for "vs.", just series seems to do the trick)


It looks incredible. I think Google are right to be concerned. Others should be too - take the mortgage example: it wouldn't be hard to integrate current deals on offer from mortgage providers.

One query I'd like to build is a TCO for a particular model of car based on running cost, service cost and the probability I will die in a crash which obviously limits my future earnings.


The success of WA is going to depend on the comprehensiveness of the data sets. If the user goes there and wants an answer and the topic of interest isn't in their dataset, then the user will revert to Google and wade through the results and read text to find the answers.

It doesn't appear to me that WA is searching text based information at all, rather these are tabular datasets that are carefully organized and queried to provide the user information.

Google is just looking for strings of text and especially strings of text from pages that have conformed to its requirements for SEO-ness leading to mostly marketing type information, where as WA is looking for facts and numbers.

I think most of the world doesn't really care about the kind of results they'll get from WA, but I don't care about that.

I think WA is a beautiful product. It isn't a google killer. Only those without a valid understanding of information would call it a google killer.

And Google Squared has nothing on WA. G^2 is just an attempt to steal WA's thunder. If Google would go to such extremes to hinder WA with their G^2, what does that say about Google's confidence in their product?

For too long, google has ridden on a history of good search results, but really, Google isn't much of an innovator. Most of their innovations were actually acquisitions. Eric Schmidt himself has said Google is not an innovator.

Here, WA is the innovator. Google is a me too and has been since their inception.


The success of WA is going to depend on the comprehensiveness of the data sets.

If you are correct, then WA can only get better and better over time. (Barring total incompetence.) So long as they can make good use of the failed queries, the user-base itself will guide what they include next.

Here, WA is the innovator. Google is a me too and has been since their inception.

The implication here that Google is the new Microsoft.


This is contrarian BS. In terms of SERP quality it was like discovering another dimension when web pages started being assigned value based on incoming links vs. on-page content. It might have been all acquisitions after they got rich off that, but it was a huge deal. I don't think it can be overstated in the terms of how it's affected the internet, it changed everything.


It gave you a job!



G^2 is just an attempt to steal WA's thunder.

Come on, Wolfram Alpha was only announced a few months ago. G^2 was not developed in a few months. It's not like G^2 is a "response" to WA.


It's way too early to be showing people the product. It is the timing of the demo that important.


The actual blog post at http://blog.wolframalpha.com/2009/05/12/going-live-and-webca... says, "We’ll be making our first attempt to go live with Wolfram|Alpha this Friday evening, May 15. We’ll start webcasting our preparations at 7pm CDT (UTC -5)."

I don't know how I feel about a Friday night launch. On the one hand the weekend allows for fixing problems, but on the other hand they are doing a webcast and have been building publicity for more than a month.


Tuesday is the best marketing day for everything.


Except everyone thinks that, which means there is more competition for attention on Tues than any other day.


why?


Saturdays and Sundays are dead.

Fridays are almost bad as nobody will remember you next monday.

Thursdays gives you just one more day to digest your offering.

Now you have mondays, tuesdays or wednesdays left.

Mondays are catch-up days, everybody getting up to the minute with meetings, planning, late tasks, errands, etc.

Wednesdays are good, very good, but no better than tuesdays, which gives you three more days for penetration, attention, conversation and full digestion.

So, even if wednesdays are good, tuesdays are better.


He will probably make a ton of money implementing this sort of system for corporations and their internal data.


Excited to see it in action and get a chance to play with it myself!


I'm betting we'll see big-time video streaming fail when this happens. The people planning these events seem always to underestimate the bandwidth that is needed.


I'm betting we won't. Justin.TV has supported more than 700,000 simultaneous viewers already. As cool as I think Wolfram Alpha is, I don't think the launch will get that many viewers.


I think he's referring to the bandwidth needed on the broadcast side. It's not much, 300 Kbps up will do, but there are plenty of times where people do these events on shared connections etc. Always frustrating.


Ok, that's a good point - it's always frustrating for us when people try to broadcast on a connection that really can't handle it.


Big fail going with Justin.tv. A lot of us refuse to use it because they still insist on spammy popups :(

A coup for justin.tv though.


Spammy popups? Definitely not to my knowledge. If anyone ever sees anything like this, please do take a screen-grab and email it to me - [email protected]


God, I love HN


Whenever I visit justin.tv I get a "blocking a popup from justin.tv" notice on FF bar. At work w/o the popup blocker it opens a poker popup.

I seem to recall it is just some pages though - I will dig out a screeny.

EDIT: voila: http://screencast.com/t/rP72HQ0UG


Popups? You mean video overlay ads?


I don't like the idea of a web cast for something like this, but if I were going to do one, I would use something like Hulu which is all the rage right now anyway. It is marketing, and Hulu has been getting a lot more air time than Justin.tv


you sure you don't have a virus on your computer? most viruses will put spammy popups on other people's web pages.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: