meta-roj

from the university of washington, i direct your attention to knowitall.

To address the problem of accumulating large collections of facts, we are developing KnowItAll — a domain-independent system that extracts massive amounts of information from the web in an automated, open-ended manner.

web-scraper fact-aggregator widget, with funding from google and darpa.

of course, what they don’t address in the papers and website is the huge amount of misinformation, disinformation and outright garbage that is harbored on the web… a problem for another day, perhaps.

this is something in which i had a miniscule contribution, so it’s time to congratulate jason scott on the bbs documentary. via slashdot, apparently the project has reached the major milestone of “the wrap.”

[update: title changed to limit spam targeting]

as covered in the washington post, aol has been blocking sites promoted via spam. it’s a controversial approach. thoughts?

from the cbc (i think – in any case, it was a radio program) comes news [in the new scientist] of the house-building robot of the future. 2005, maybe. behrokh “berok” khoshnevis and yan xiao brings us this wonder. and it’s well on its way – from what i’ve seen 2005 isn’t just vapor – even if they don’t quite get all the painting done, it’ll still be a good chunk of a finished house.

this could be a big change for habitat for humanity. it could also be a lot bigger than exporting geek jobs to india.

a random statistic i pulled out of the thin web (credited to the building industry association) suggests that it takes 1100 human-hours to build a typical house. another random statistic says that there were 1,700,000 housing starts in 2004. that’s about 1.87 billion human-hours devoted to building houses in the united states (close enough for blog work). so when this thing does its thing, then this housing-construction market could collapse (we shared something vaguely like this before. 1.87 billion human-hours (that’s 935,000 2000-hour work-years)… say… 408 million (that’s three 8-hour shifts for 10 days, or 240 hours per house).

i’m just speculating, of course. eventually, maybe you just drop this machine off on a worksite with some architectural plans and show up every couple hours with a new truckload of building-goo. given our experience with inkjet printers, i think you can count on paying more for the building-goo (in special, drm-controlled, chipped, patented containers) than for the machine.

then we have to revisit the concept of “homeless in america”

update (2005.04.10): this story now covered in discover and subsequently slashdot

this just flashed through cnn as a teaser, and cbs has a story about people watching porn in… *gasp* vehicles. mobile porn. wow.

[cue paradise by the dashboard light]

is it really a surprise? the looney tunes edition suv isn’t just for cartoons anymore… [insert reference to technological determinism]

and it’s not really new news, a few weeks ago, a similar story was picked up as far away as south africa and new zealand (or maybe it’s just something about reading about watching porn in the southern hemisphere…)

so i got clued in to a little tempest about blogs and such…. and i’ve been meaning to ask questions about the ability to engineer emergence, but for now, i’ll just poke my fingers in a few obvious eyes 🙂

there’s a broadcast in bottom-up clothing rant, and the thing that started it (or at least this round of it)… and a few other things out there…

well, fear not, here at meta-roj we’ve been on this case for a long, long [internet] time… way way ahead of the curve.

in the meantime, i’ll just ponder the thought that we’re all little experiments to lots of things (people, companies, governments, whatever…).

i picked this up from the ap wire, but it’s something i’ve been flirting with for a while… i just didn’t have a good hook until i read this:

vivisimo helps narrow internet searches

Let’s say, for example, you’re curious about accommodations in France and enter a search for “Paris Hilton.”

Google recognizes this as a search in the category of “Regional-Europe-Travel and Tourism-Lodging-Hotels” but still produces page after page with links about celebrity socialite Paris Hilton and her exploits. That’s because Google’s engine ranks pages largely based on how many other sites link to them, sending the most popular pages to the top.

(i’m just going to give them the benefit of the doubt that someone looking for paris hilton on google has any interest at all in france).

i’ve been watching categorization/searching/organizing applications for a few years now, and one of the more interesting ones a few years ago was from a local company (wisdombuilder). i don’t know if they’ve kept up or done anything really new and exciting, but it was neat to play with for a while, particularly since i had a particular application. that gig didn’t work out. but, i digress. you people come here to get interesting perspectives, so here’s the line you want: the paris hilton problem is the killer app in this category.

in a few years, when google is all grown up, we’ll be talking about the day that google passed the “hilton test.”

that, of course, will be a few weeks before google passes the “turing test.”

update: kevin pointed me to this at ieee spectrum. now that i’ve defined the hilton test (i shoulda used my own name), the race is on…

update (2004.01.05): slashdot picked it up.

i just have to stick my nose in this one… i just saw jeremy zawodny (missing the point of social networks) throwing a $1000 gauntlet into the net, and within 3 hours or so, it’s accepted by richard stokes (i’ll take social software for $1,000 please, alex).

my name isn’t alex, but i just have to see how this plays out. i suggest our two friendly dissenters hie themselves to long bets and make it official. we have a slight gap between the two bets at this point,

jeremy zawodny

If you really think that Friendster scored over $10 million in VC funding on the premise of being nothing more than a “browse your friends” site, get real.

Or maybe I’m the one smoking crack. But I’d bet you $1,000 I’m not.

richard stokes

Given the historical track record that 70% of all VC-backed ventures go nowhere (or worse), I’ll be happy to take an even-money bet for $1,000. How do you propose we decide if it is a commercial success or not?

so we’re not exactly on the same page with “commercial success” versus “nothing more than a browse your friends site,” but there’s got to be a way to make this work. surely, keeping me entertained in some small way is worth $1000 to these two gentlemen :)….

since we’ve got the vc angle, i suggest a short-term long bet, something on the order of 3 to 5 years, to match the vc horizon. this also needs a fundamental metric of some sort, such as “still operating.” from there, jeremy and richard need to work out what constitutes a significant change in the friendster model.

these are smart people. this should be good.

update (2004.01.06): updated link for startupskills

i’m allergic to powerpoint. there are debates on the issue, and i’ll leave those to the microsoft pr department and edward tufte, since they’re much more qualified to make their individual cases.

personally, i just sat through a few too many powerpoint presentations where all the speaker did was read the screen to me. this, i can do on my own. this, i can do faster on my own. this, i can do at 3am. so i don’t much care for powerpoint. especially busy, animated powerpoint.

then david byrne goes and picks up powerpoint.

there’s some potential for redemption here.

[text of the abc piece below]
(more…)

rob has some interesting company – while channel-surfing, i just learned that neither larry king nor oprah winfrey have cell phones.

i guess rob is cooler than he thought 🙂