It might be easy to dismiss this as a straw man, with the spurious creation of `@[i]` every loop but there's a more realistic gotcha here.
between for i in 1..1e5: result &= i Run and for i in 1e5: result = result & i Run The first appends to result, and is fast, the second performs many allocations and is slow. I suppose the second could be auto optimized into the first, maybe even with a term rewriting macro but I smell edge cases here I don't know