This is why I prefer dating contracts to deep technical interviews. If someone shows enough promise in their CV and portfolio, and they're personable enough to sell themselves reasonably, then a 1-2 week contract to cover some real work is low risk and a better measure than trivial whiteboard problems.