Chat Absolutism

From the AI Research Compass wiki

Chat Absolutism: The chat box is the one true interface to AI, and every capability should be delivered as a turn in a conversation.

Beliefs

Chat Absolutism holds that the conversational turn is the atomic unit of human-AI interaction. User types, model replies, repeat. Search, writing, coding, tutoring, analysis and customer service all reduce to a dialogue, and anything that cannot be expressed as one is either a niche or a design failure. Dashboards, buttons and autonomous background agents are distractions from the purest interface ever invented: a text field and a send button.

It sits left of centre in the upper half of the chart: language-first because dialogue is made of words, and Monolith-leaning because the doctrine centers one model answering one conversation rather than a system of agents working on its own. Its paradigm level is orthodox. The chat interface defined the first wave of mass-market language models, and the next wave is increasingly measured by how far it moves beyond it.

Arguments

  • Training models to follow instructions in dialogue made them dramatically more useful and accessible than raw completion models.
  • Conversation is the most universal interface people already know, requiring no training or documentation.
  • Multi-turn dialogue naturally supports clarification, correction and iterative refinement.

Critiques

  • Many valuable tasks are long-running and autonomous; forcing them into synchronous turns wastes both human and model time.
  • A text box hides capabilities and offers poor affordances for structured work such as spreadsheets or design.
  • Agent and computer-use advocates argue that acting in software matters more than talking about it.

Representative works

Relationships

How to get it

Land on this cell in the full or pro test: x ≈ −0.66, y ≈ +0.62, paradigm ≈ −0.76.

A first draft. Corrections and better arguments are welcome through a pull request. Edit this entry on GitHub Edit its data