>And that's a shame because the live coding exercise is both a better predictor of hands-on job success and it's a better recruiting tactic

I'm very interested in this claim. Every interviewing-is-broken thread, someone mentions that pairing, or take-home exercises, or work samples, or their favorite method, are better predictors than algorithm problems. On the one hand I doubt any but the biggest companies could conduct studies on this; on the other hand individuals at smaller companies might honestly remember only those times their hires filtered through this method were successful, due to confirmation bias. So it seems like an inherently difficult thing to study.

Whereas, algorithm questions lend themselves easily to a "rubric" and seeing, 1 or 2 years later, if higher "grades" corresponded to better job performance.