Comments on “Run-time task-relevant algorithmic understanding”