Where interaction becomes structure.
Long-form human-LLM interaction, examined as structure rather than content. Across hundreds of turns, interaction develops patterns, variation, and measurable structure that short exchanges cannot reveal.
Why long-form interaction.
Most of what we know about language models comes from short exchanges — benchmarks, single prompts, brief task completions. The interactions that matter in practice are longer: research threads, working relationships, extended analysis. At that scale, something different becomes measurable. Pacing, coupling, and structural variation emerge that aren't visible in a single turn and cannot be inferred from one.
We treat long-form interaction as its own research object — not a longer version of short interaction, but a distinct regime with its own dynamics worth measuring carefully.
Collaboration
Available for selective research and advisory partnerships.
For collaborators who need careful analysis, rigorous framing, and interpretive depth across research, evaluation, and applied work.
Contact Axiom Drift AI