The AI ecosystem is becoming confusing even for those who follow it daily

Modes are multiplying faster than their map

Ethan Mollick described the confusion created by the proliferation of AI modes and interfaces. ChatGPT Work, Codex, and Chat now sit alongside Claude Cowork, Code, and Chat, with further divisions between desktop and web apps. Even someone who follows the field every day may no longer know exactly what belongs where and how the capabilities differ.

Switching tools means searching for context again

The uncertainty extends beyond product names. Mollick asks which mode contains plugins, skills, memory, permissions, and files. When users move between chat, a work mode, and a coding tool, they must determine again whether the same resources and context remain available, or if they are starting with a blank slate.

Invisible boundaries raise the cost of use

Each additional mode may serve a purpose, but without a clear map it adds another decision to the work. Users have to manage not only the task itself but also the choice of interface and access to data. The problem is therefore not the breadth of the offering alone, but boundaries that are difficult to see and that slow down daily routines.

Shared data visibility is a necessary condition

More important than another new mode will be whether providers clearly show which plugins, skills, memories, permissions, and files carry across applications. The visibility of those connections will reveal whether the AI ecosystem can keep expanding without creating further frustrating confusion.

Lilith's verdict

Humans wanted tools that save time, and vendors gave them a maze full of unmarked doors and siloed memories instead of a clear map.