If there's a singularity in the next decades, all bets are off anyway and Alphabet's CFO can't plan for that. In the unlikely event that there is not, it means AI probably hit a plateau. In the double unlikely event that AI takes off yet there's no singularity and the economy remains recognizable, then we are in for decades of double digit real rates. Selling very long dated bonds is a good idea. I'm not sure why they should plan for the third option over the second though.

Currency would debase so might as well max debt now