The better way existed before AI existed. Take-home tasks were always poor attempts to find a measure for technical ability. Pairing exercises or live problem-solving tasks were, and still are, vastly more effective. If your team uses AI for those challenges on the job, then the people you interview should too.
The measure should always be the quality of a solution a candidate produces in response to a problem, not how they got to the solution. If you want to dive into the 'how' then what you're actually measuring is communication, something AI is still pretty bad at.
The measure should always be the quality of a solution a candidate produces in response to a problem, not how they got to the solution. If you want to dive into the 'how' then what you're actually measuring is communication, something AI is still pretty bad at.