There was this Australian company that had this fly-by-night looking empty office near South Park, San Francisco... I'm serious, and not exaggerating, when I say it was a few bare metal folding chairs/tables bare walls, and way too much open space for the 1/2 dozen people working there...
In any case, the startup I worked for previous was split up and sold/aqui-hired out to two different companies, and myself and the technology I'd created, and the million+ email addresses we'd collected went to this one.
I went along with it because they were one of the few companies I knew of with access to the twitter fire hose feed, which interested me, and they had dept. of defense contacts - which ( in my mind ) legitimized their presence in the U.S.A. at the time.
Out of the many shady-feeling project, the one that made my skin crawl the most was an automated fuzzy matching system I was ordered to create, that tried its best to match users on Twitter to users on Facebook, using nothing beyond the normal publicly available meta data ( name, age, icon, the regular profile stuff ) and the followers/people followed on the twitter side, and friends and likes on the Facebook side.
It was surprisingly easy to match people, and felt more than a little wrong to me when I really thought about it... but the tech/challenge was just too fun to work on to routinely give it more than a perfunctory thought.
In any case, the startup I worked for previous was split up and sold/aqui-hired out to two different companies, and myself and the technology I'd created, and the million+ email addresses we'd collected went to this one.
I went along with it because they were one of the few companies I knew of with access to the twitter fire hose feed, which interested me, and they had dept. of defense contacts - which ( in my mind ) legitimized their presence in the U.S.A. at the time.
Out of the many shady-feeling project, the one that made my skin crawl the most was an automated fuzzy matching system I was ordered to create, that tried its best to match users on Twitter to users on Facebook, using nothing beyond the normal publicly available meta data ( name, age, icon, the regular profile stuff ) and the followers/people followed on the twitter side, and friends and likes on the Facebook side.
It was surprisingly easy to match people, and felt more than a little wrong to me when I really thought about it... but the tech/challenge was just too fun to work on to routinely give it more than a perfunctory thought.