LLMs Are Mathematically Equivalent to a Book, Argues Imbue Cofounder in Consciousness Debate
joshalbrecht · x · 2026-10-03
Josh Albrecht (Imbue) pushed back on the growing AI consciousness debate with a crisp argument: an LLM is mathematically equivalent to a book — and nobody thinks books are conscious.
The logic:
- An LLM is a pure function: f(w, s, t) -> o (weights, seed + settings, input tokens → output tokens), with no side effects.
- Since inputs and outputs are fixed-size, any such function can be rewritten as an explicit lookup table — literally a printed book mapping every input to its output.
- Hence LLMs fall in the class f(ℕ) → ℕ, book-like by construction. QED.
Caveats he notes:
- The argument targets LLMs only; agents (tool use, execution over time) could in principle be a different story.
- Context: a Steve Yegge post predicting consciousness debates will dominate discourse next year as "AI employees" proliferate — Yegge, after fifty years reading philosophers and holy works, claims no one has ever understood consciousness and offers a simple rule to pick a side.
More from AGI Musings
- Reading AI slop costs an order of magnitude more energy than generating it — sebpaquet · 2026-10-03
- The EinsteinTest: models given pre-breakthrough evidence rarely make the break themselves — CurieuxExplorer · 2026-10-03
- Reddit argument: AGI may emerge from system architecture, not one frontier lab's giant model — sourdub · 2026-10-03
- "Software has become liquid": Reddit on SOTA models dissolving the walls between programs — a300a300 · 2026-10-03
- Game studio CEO Irena Pereira: games must be made by people, not AI — CurieuxExplorer · 2026-10-03
- Signal president Meredith Whittaker on defending art and artists in the age of AI — mer__edith · 2026-10-03