Google Squared is an Exponential Improvement in Search

If the exponent is negative, maybe.

But, let's not blame Google for that hype. They haven't claimed this is anything other than an interesting experiment.

At the moment pretty much all Squares are less useful than a straight web search. Perhaps that will change as they acquire more data, but I'm skeptical. As a way of accelerating a human's ability to organize data, maybe it has a brighter future.

It's a toy, no more, no less.

According to Google Squared, our Russian President has been dead for a couple of weeks by now (he's alive and well, just in case you're wondering), we have been also ruled by someone named "5 November 1996" and "Constitution" was our President too. The one guy left in the list is not a president at all (Viktor Chernomyrdin is Russia's embassador to Ukraine). http://www.google.com/squared/search?q=presidents+of+russian...

Exponential improvement, eh?..

"exponential" might be an attention-grabbing stretch, but it is clear that we will discover new uses for this.

I am wondering if anyone knows the relative size of the data behind wolfram alpha compared to google. I know that wolfram uses "trusted" sources for its data.

I am recalling a recently quoted paper by Norvig that seems to imply that having multiple of orders of magnitude of data available changes the knowledge engine game.

It's a great way to organize and display search results in certain situations, but I wouldn't call it an 'exponential improvement in search' just yet. Getting the actual data (i.e. the search) leaves a lot to be desired. It's the format itself that I think is a great idea.

It's nowhere near useful.

It's a nice tech demo, but nothing more. I doubt it will ever be usefull, because that would require real semantic data or real AI (and neither seems to be around soon).

It's definitely not production ready, but I still hold out hope. Two things right off the bat to increase its functionality would be to make the columns sortable and to provide ability to set the column headers equal to a particular value.

For example, if I run a search on San Diego restaurants to get an idea of where to go for lunch, I'd like to be able to set cuisine and price and build from there. I could see other comparisons working similarly.

Defiantly not anything useful at the moment, even there example searches seemed a bit sketchy.

I think it will probably be a company like Google that cracks this problem just because of the sheer data/ processing power you would need available to have a good crack at it, we are a fair way off that though.

It's a neat thing, but wow, this would be over-hype even if it worked much better.

RDFa, we need you now!