Gpt-4 - What GPT-4 Cannot do?

Neither hearing nor speech are capabilities that GPT-4 possesses. The 'Whisper' model uses GPT-4 to analyze text that has been transcribed from spoken input. Another model is then fed the results. The problem is that this approach can't pick up on nuances in speech such as accent, tone, emotional expression, and pronunciation.