Ahem.
I'll put up my hand as a completely non-technical editor who reads this list on a regular basis, and who shares Ryan's "eye-bleeding" feelings about templates as they are currently developed and utilized, at least on enwiki. I think I can speak for a very large number of editors who have been praying for years for "something" to improve the template use experience, and say thanks to the development team for answering the call and coming up with what seems like a pretty good solution.
I quite agree that it's important to actively seek out and work with template creators and curators to get involved in testing out Lua. But it is also critical to involve those who will be *using* the templates to see if what has been designed will address any or all of their concerns and issues. Making the wiki run faster is, I believe, only one objective in improving the "template experience". If the product that comes out of Lua is as difficult to use as the current templates, then we have only completed half the job. (I use "we" as the big-tent community that includes developers and editors.)
If I may suggest: it's probably not all that hard to identify which projects use a lot of templates (enwiki will always be one), and then find a dozen or so template creators/curators from those projects - ask them to create the most commonly used templates on their project using Lua. Then, ask a larger group of template *users* come and try them out to get the end-user experience. This should give some useful feedback to Tim and anyone working with him. It also has some serious potential to create some evangelists within the project who can build up the positives and start moving their respective communities toward embracing Lua. I don't know if mediawiki is the best place for this - perhaps a test wiki might be easier to deal with, since the Mediawiki community has a very different ethos than every other project - but I think it will be worth the investment.
If ever there was a development project that needs a dedicated ambassador or two, this is the one. It has the potential to significantly affect multiple facets of the user experience. It has the potential to be a very big "win" for everyone - developers, editors, and even readers.
Risker/Anne