The article discusses new job roles that may emerge due to the rise of machine learning systems, such as 'incanters' who optimize interaction with language models, 'process engineers' for quality control in legal documents, 'statistical engineers' to manage variability in ML outputs, and 'model trainers' who improve model accuracy with expert input. It also mentions the need for 'meat shields' to take accountability for ML failures and 'haruspices' to analyze model behavior when things go wrong. The author reflects on the future of work alongside these transformations, while addressing the implications of AI on accountability and employment.