Currently associative destructuring emits calls to get. The attached patch modify desctruture to emit direct keyword lookups when possible.
Rethinking about this patch now, it may be too specific: get's inline expansion should be modified when the key is a literal keyword.
More generic patch (inline-get-keyword.diff): all get calls with literal keywords as keys are inlined to direct keyword lookup.
Is this only stalled out of lack of interest?
There are currently about 50 tickets "triaged", i.e. marked for Rich to look at and decide whether they are things he is interested in seeing a patch for, and another 25 or so that were triaged and he has "vetted" them, and they are in various stages of having patches written for them, screened, etc. That doesn't mean anything for this ticket in particular – just wanted to make it clear that there are a bunch of other tickets that are getting some attention, and a bunch of others that are not.
What gets triaged depends somewhat upon how severe the issue appears. You can vote on the ticket, and try to persuade others to do so as well, if they think this would enhance the performance of some commonly-written types of Clojure code. You could also consider doing some benchmarking with & without these patches to see how much performance they can gain.