Reproduce “Talking with Language Models” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
James Ravi Kirkpatrick, Alexandru Radulescu, Rachel Katharine Sterken
The paper argues current LLM interaction is better understood as candidate exchange than conversation: users prompt systems that generate text, then interpret and take it up. The authors say fluent outputs alone do not establish a speaker or commitments, while users supply framing and uptake. This conceptual argument offers no empirical evidence that conversation is impossible.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.