The strength of the triple quad is to use prior knowledge of how peptides (in my case) fragment to detect them reproducibly in complex mixtures. This is a critical and non-trivial step. I am using Skyline - a wonderful open source program developed by the MacCoss Lab - to mine our extensive collection of peptide spectra collected on our Orbitrap and develop methods for the TQS.
So far this is working pretty well; many transitions from the standard UPS2 mixture have behaved. Stright off the bat I am down to 500 attomoles on column. When I optimise the spray position more, change to a better column, define retention tiems and play with collision energies I think 5 attomoles is within our reach.
Things have gone so well that I am now looking at a 'real' protein and trying to have a quick look at some phosphopeptides before I go on holiday.
[edit Apil 2012 - actually 5 attomoles on column is still below my reach. 100 attomoles, yes if everything is in tip-top condition (clean but not absorbant column, clean source, stable spray, no stray polymers...oh yes and nice tight peaks - no post-column dead volume!)]