The model does not. The system it runs in can. My agents have persistent memory. It is not at all clear whether or the distinction that the long-term memory is external from the model matters.

The 'static model' framing is mostly right but does some unintended work in the consciousness debate.

Human neurons aren't static between thoughts -- they're stochastic, not deterministic. Firing isn't guaranteed; it's probabilistic. What gives humans consistency isn't a frozen substrate but a persistent memory architecture that stitches moments together.

The more interesting question this framing opens up: what's the temporal grain of any potential LLM experience?

Human consciousness isn't a continuous stream -- inattentional blindness, change blindness, and the binding problem all suggest it's discrete moments stitched together. A transformer forward pass has a definite temporal boundary. If there's anything it's like to be an LLM, the question 'during what?' matters enormously. During a token? A context window? A session? These aren't equivalent, and none of our existing frameworks for consciousness were designed to address it.

Nagel's 1974 'What Is It Like to Be a Bat?' is still the clearest framing. His point wasn't that bats definitely are conscious -- it's that some experiences might be genuinely incommensurable with human experience. Not more, not less, just different and untranslatable. Every test for LLM consciousness we might propose was built to detect human-like consciousness. That may be the wrong instrument entirely.

It's different, but it is not at all clear whether or not that matters for the purposes of whether or not they can experience things, seeing as we don't know how to measure that even for other humans other than by relying on self-reporting to infer without other evidence that their experience matches ours.