One has to be careful about exit polls. In the 1992 UK general election, an exit poll of 10% of the voters turned out to predict the wrong result by more than could be explained by pure statistical chance. It turned out that a significant proportion of the electorate had "forgotten" which way they voted and said that they'd voted for the other party. It was an interesting evening watching the BBC's swingometer swing past the margin of error as the actualy results came in (being a country with a high population density and a simple voting system means we get the first fully counted official result within about 45 minutes and it's only if there are tight recounts that there are any results going much past breakfast the next day). Needless to say the exit polls now include a weighting factor for "too embarassed to admit who you voted for..."

Ok, so I know that a magnetic disc around 100-200nm in diameter will have a vortex domain structure (actually - I don't magnetic nano-rings tend to form onion states, but we'll leave that one for now). I know I can set the vortex state by hitting it with a high intensity laser circularly polarised laser pulse and apply a small +-z magnetic field to set the in/out state. But reading it needs some fancy focussed magneto-optical kerr effect kit and a lot of patience and re-writing the data is tricky - you have to de-magnetise the nano-disc and start over.

Seems a lot of effort for only a factor two increase and a technology that still requires moving parts to move the data element and read head into alignment.