In the last two years I've gone from thinking the simulation hypothesis is probably false to almost certainly true