It's currently dependent on implementing traits for very large tuples, so you could argue it's a language feature we are waiting for (cf. open RFC #1935 [1]). (There was a PR on Diesel to use HLists but it's no longer pursuit.)
What was the reason you guys dropped HList approach? I'm curious since I have to deal with tables with north of 200 hundred columns and because of forever annoying Scala's tuple parity limit (22) Slick falls back to HList which is really heavy on a compiler and IDE aside from other drawbacks.
We didn't want to commit to our own home-grown hlist type until we had some indicator of whether the language was going to go with variadics based on tuples or hlists.