Jason Calacanis has
launched Mahalo in alpha.....its a "Human powered" search engine that aims to only search the "fat head" of search topics, and aims to have the top 10,000 or so in by year end. As
Valleywag put it:
Points for a contrarian approach: the orthodoxy is that only machine indexing of the web can satisfy users' obscure curiosity; Calacanis says that doesn't matter because the top 10,000 search terms account for about a quarter of all queries. They're also the terms against which the majority of text advertising is bought.
It is interesting....Google is observably getting less useful as it gets more gamed (and as the web simply gets richer in content) so humans can surely add value - but using humans rather than algorithms means it is far harder to scale the operation, so this probably forces the economics of Pareto (the 80/20 or "fat head") on Mahalo, rather than that of the Long Tail.
However, it is probably also reflecting the reality of online Ads today - the Long Tail may be great for flogging stuff directly, but no Adman wants to pay anything interesting for that eyeball in the long grass. In theory interactive advertising will change all that, but today the value is still in volume - - probably more out of history than economic reality, but things take time to change.
But, will this business scale - rather than sale

- as its not hard to imagine doing the same 80/20 job with a far more automated service and still hoovering up the Ad cents. Also, will people really be happy to continually go to a site that doesn't search everything and anything?