Model5d ago
Language Models: Inside Their Minds
A thread on Lobste.rs drew 1,200 comments, asking what it feels like to be a language model. Participants cited OpenAI’s GPT-4, noting its token‑by‑token processing and lack of consciousness. The debate shows that developers can’t assume a model’s internal state is human‑like. Practitioners should treat models as deterministic, data‑driven systems.