I wonder, could it make sense to build houses in risky areas more like ships?
Sure for a really big quake it would be more significant but the engineering there is really impressive.
He's lived there quite a while and has experienced several which have woken him up.
San Francisco's last slate-cleaning came in 1906, though parts of the city were constructed far more recently. The 1989 earthquake (nearly 100 km distant and comparatively mild (6.9, as opposed to an 8 or 9 mag such as the 1906 quake) still caused significant structural damage throughout San Francisco, though major damage was limited to a few neighborhoods and specific buildings -- more major damage was concentrated largely to the south, closer to the epicenter.
Most of Tokyo was destroyed during WWII, and has been rebuilt since. Japan also experiences frequent powerful earthquakes (several M8+ quakes are noted by the USGS: http://earthquake.usgs.gov/earthquakes/world/historical_coun...), effectively requiring more robust building standards. Though my understanding is that there's a great deal of collaboration between Japanese and Californian engineers on this work.
And yes, M9s are monsters -- the Boxing Day and March 11 earthquakes in Indonesia and Japan were both M9s, I believe there was one in Chile ... ok, that was 1960 -- the 1964 Prince William Sound, Alaska, quake was also 9+. Listing at Wikipedia: https://en.wikipedia.org/wiki/List_of_earthquakes
On the US West Coast, it's the Oregon / Washington coast which would be most likely to see a quake of that magnitude, along the Cascadia fault, a megathrust fault which could produce similarly large tsunamis as well as massive shaking. For most of California, not so much.