Friday, February 06, 2004
Mike Perkowitz responded to my questions about using decayed influence to weight the value of a friend's influence on my search. Link.
Peter, the idea about decayed influence just seemed a natural. I haven't heard it applied to the social network question before, but you see similar ideas in other contexts, for example in discounting future positive/negative feedback, or in the influence of more and more distant events on probabilities. Mathematically, the natural thing would be to have influence decay exponentially (just as your network size presumably expands exponentially with each hop), but that might decay too quickly, resulting in a social tool that doesn't do anything!
Sounds like we need an actuary!
Gramercy Strategies is Eurekster's communications firm. Looks like Gramercy's premier client is NBC/NBCi and is based in NYC. I guess this shows a bit of the pull that NBC has over Eurekster. Link
Someone wrote about the eureksterblog in German and I don't know a lick. A translation would be appreciated. Link
Some nice kudos about the eureksterblog from Marcus P. Zillman, an internet jack-of-all-trades: author, speaker and consultant. Thanks Marcus. If you haven't read his blog, check it out. Link