It is closer to fair sampling than you think, since the average internet user is more likely to install toolbars unlike the technogeek niches that we hang out in. Also, it's more useful for relative changes like drops and increases rather than absolute percentages.
Also, it's one of the only few samplings we have, unfortunately.
I think you missed my reference to your specific point.
"No-one I know - from the savviest sysadmin who only browses in lynx to the most malware-infested bloated-PC user - has ever used Alexa toolbar."
I'm one of "those guys" who everyone comes to in order to get their shitty computer fixed for free. I've seen a lot of "normal users". I do not think that Alexa users are a representative sample based on what I've seen.
Also, it's one of the only few samplings we have, unfortunately.