Designer and developer Nikita Prokopov published a detailed critique of menu and toolbar icons in macOS Tahoe, arguing that Apple’s system falls short of the basic purpose of iconography: helping people locate and understand actions quickly. His assessment draws on screenshots from stock Apple applications in macOS 26.1 and 26.2, with no modified system settings.
Prokopov’s first objection is density. When nearly every menu command receives a symbol, he argues, no individual action stands out. In his view, monochrome presentation further reduces visual separation between labels and their icons. He acknowledges that color would require more deliberate design and might not look immediately appealing, but maintains that stronger differentiation would improve scanning.
Consistency forms the core of the critique. The post collects multiple symbols used for actions described as “New,” along with varying treatments for common commands such as Find, Search and Filter. It also compares toolbar and menu symbols inside the same applications, showing cases where conceptually similar operations receive different marks. The reverse problem appears too: identical or nearly identical icons are assigned to different actions, sometimes within a single menu.
That reuse matters because an icon becomes useful when a person can learn its meaning and recognize it elsewhere. Prokopov argues that Tahoe weakens that learning process. Small variations in arrows, pencils, document sheets and three-dot marks are expected to communicate different functions, while Apple also uses visibly different versions interchangeably in other contexts. The result, in his account, is a system that asks users to notice fine distinctions without applying those distinctions reliably.
Physical size is another concern. Prokopov says many Tahoe menu icons fit within a 12-by-12-point space, rendered as 24-by-24 Retina pixels. He compares that with the 16-by-16-pixel icons used in Windows 95 and calculates that, at representative display densities, Tahoe’s symbols occupy a substantially smaller physical area. His examples highlight details such as tiny letters, window controls and line-weight changes that may be difficult to discern at normal viewing size.
The post attributes part of the rendering problem to vector font symbols. That approach lets Apple reuse artwork across sizes, resolutions and weights, but Prokopov argues it complicates vertical placement, stroke alignment and mapping onto the pixel grid.
This is a design critique, not a usability study, and the supplied evidence does not include Apple’s response or controlled performance measurements. Its contribution is a first-principles case supported by many interface examples: icon systems work through salience, stable meaning and recognition at a glance, and Prokopov contends that Tahoe’s implementation compromises all three.
