That he thinks the argument is sound but does not agree with "(3) we are almost certainly living in a computer simulation" means Bostrom believes either:
(1) the human species is very likely to go extinct before reaching a "posthuman" stage; or
(2) any posthuman civilization is extremely unlikely to run a significant number of simulations of their evolutionary history (or variations thereof).
In the FAQ, Bostrom says he personally assigns about 20% probability to the idea we are living in a simulation, and that he views (2), that the posthumans are unlikely to run ancestor simulations, a bit more likely than the others.
Colloquially, it would be fair to say he 'believes' (2) -- as in the most likely of exclusive options, if he had to make one choice, if he had to bet, etc. And he definitely believes Union(2+3) more than (1).
Do you really believe that we are in a computer simulation?
No. I believe that the simulation argument is sound.