The article reviews recent advances in understanding working memory, challenging the traditional view that it has a fixed capacity limit. It argues that working memory should be conceptualized as a flexible, limited resource distributed among all items to be maintained. This resource-based model explains why performance declines with increasing set size and improves with salience or goal relevance of stimuli. Behavioral and neural evidence supports this view, showing that recall precision decreases continuously with set size and that neural activity in prefrontal and parietal cortex reflects this continuous resource allocation. The article also discusses the implications of these findings for neural models and future research directions, emphasizing the need to consider both the quantity and quality of representations in working memory.The article reviews recent advances in understanding working memory, challenging the traditional view that it has a fixed capacity limit. It argues that working memory should be conceptualized as a flexible, limited resource distributed among all items to be maintained. This resource-based model explains why performance declines with increasing set size and improves with salience or goal relevance of stimuli. Behavioral and neural evidence supports this view, showing that recall precision decreases continuously with set size and that neural activity in prefrontal and parietal cortex reflects this continuous resource allocation. The article also discusses the implications of these findings for neural models and future research directions, emphasizing the need to consider both the quantity and quality of representations in working memory.