I would just take this to mean that most people are bad at estimating timelines for complex engineering tasks. 15 years isn't a ton of time, and the progress that has been made was done with pretty limited resources (compared to, say, traditional microprocessors).