A hard to attain science and engineering task is often 5 years away for a long time until 5 years before it is actually delivered. Unlike building widgets building something never done before entails an awful lot of hard to estimate work. Anyone doing software engineering should be familiar with this.

The “5 years for 10 years” meme seems to be confusing a lack of accuracy when giving a precise answer to a difficult to estimate task for either foolishness or fraud, or perhaps both. To me it reads like every software and research estimate I’ve ever seen - a sliding estimate on something you know will deliver but hard to pin down exactly all the unknowables are worked out.