I think that trying to explain all this stuff (big sweeping gesture) to people without technical backgrounds is beneficial exercise, for several reasons. Any major programming or business work is going to involve quite a bit of communication, with people as well as computers. Being unable to comfortably explain how your creation works to a potential client is a great way to get rejected, and is probably at least as damaging as being unable to be on the same page as a collaborator. Also, explaining something to someone else can help you discover aspects you don't fully understand, and often they have useful abstractions that may not have otherwise occurred to you.
His Python explanation is a bit mystifying (particularly the hand-waving about static vs. dynamic typing at the end; if he was going to just skip them, he probably could have avoided bringing them up), but to his credit, this can be really hard to do well.
For one example, I'd compare the difference between an interpreter and a compiler to an interpreter standing between two diplomats and a translator slowly rewriting poetry or legal documents. The former has more immediate feedback to work with (raised eyebrows, for starters), but has to do everything on the spot, while the latter has a lot more time to rewrite things concisely and elegantly. In this sense, an interpreting person who is quite fluent in both languages and excellent at reading body language might be comparable to a skillfully optimizing JIT-compiler.
I think mail-order packaging could be a useful metaphor for part of the efficiency difference in static vs. dynamic languages: a static language determines how much space something will need in a shipping container, while a dynamic language saves time by putting most things in roughly the same size boxes and loading them up with plastic shipping air bubbles. Granted, this is only part of a complicated issue, but it's still a useful abstraction, without getting into a bunch of details about boxed and unboxed types, type systems, pointers, and other implementation details.
His Python explanation is a bit mystifying (particularly the hand-waving about static vs. dynamic typing at the end; if he was going to just skip them, he probably could have avoided bringing them up), but to his credit, this can be really hard to do well.
For one example, I'd compare the difference between an interpreter and a compiler to an interpreter standing between two diplomats and a translator slowly rewriting poetry or legal documents. The former has more immediate feedback to work with (raised eyebrows, for starters), but has to do everything on the spot, while the latter has a lot more time to rewrite things concisely and elegantly. In this sense, an interpreting person who is quite fluent in both languages and excellent at reading body language might be comparable to a skillfully optimizing JIT-compiler.
I think mail-order packaging could be a useful metaphor for part of the efficiency difference in static vs. dynamic languages: a static language determines how much space something will need in a shipping container, while a dynamic language saves time by putting most things in roughly the same size boxes and loading them up with plastic shipping air bubbles. Granted, this is only part of a complicated issue, but it's still a useful abstraction, without getting into a bunch of details about boxed and unboxed types, type systems, pointers, and other implementation details.