Coding challenges are often time limited, but AI has created an arms race at the assessment level, as coding agents are becoming so good that they are blasting past human performance speeds leading to the need to rework the assessment or reduce te time permitted. This is true even (and maybe especially) of Anthropic, whose coding challenge is now
publicly available for anyone to try, as its now been made unfit for purpose by Anthropic’s own coding agents.