Software spent decades adding interface. More controls, more screens, more dashboards. The current direction is more interesting: mature products are beginning to remove steps.
Automation, natural-language input and predictive defaults make it possible to express an outcome instead of operating every intermediate control. The user asks for a result; the system composes the actions.
Less interface, more responsibility
This does not eliminate design. It raises the cost of bad design. A button makes its action visible before the user presses it. An automated agent may execute ten invisible actions before the user sees the consequence.
That means the new interface primitives are confirmation, reversibility, provenance and clear boundaries. Users need to know what happened, why it happened and how to undo it.
The future is not screenless. It is selective. Interfaces will remain dense where exploration matters and collapse where the intention is repetitive and well understood.