I'm curious as to how you calculate that >95%. I ask because I, personally, overestimated the threats from what amounts to unfriendly AI at two points in time (during the Japanese 5th generation computer project, and during Lenat's CYC project), and I overestimated the threat from y2k (and I thought I had a solid lower bound on its effects from unprepared sectors of the economy at the time). Might you be doing something similar?

Full disclosure: I have cryonics arrangements in place (with Alcor), but I'm unsure whether the odds of actually being revived or uploaded justify the (admittedly small) costs. Since I've signed up (around 1990 or so) I've revised my guess as to the odds downwards for a couple of reasons: (a) full Drexler/Merkle nanotech is taking much longer to be developed than I'd have guessed - "never" is still a distinct possibility (b) If we do get full nanotech, Robin Hanson's malthusian scenario of exploding upload replication looks chillingly plausible (c) During the Bush years, biodeathicists like Leon Kass actually got positions in high places. I'd anticipated that life extension might be a very hard technical problem - but not that there would be people in power actively trying to stop it.