We construct a universal Turing machine with five states, five symbols and 20 instructions in the standard model: one bi-infinite tape, left and right moves only, finite input on a blank background, and halting at an undefined transition. This improves on our earlier 21-instruction machine and other known machines in this model by Rogozhin and by Neary and Woods which use 22. The saving comes from a redesigned printing and restoration protocol that exploits a parity invariant. A direct compiler, with no intermediate tag system, drives the machine through a queue of unary addresses, periodically rebuilding its program so that the history it traverses stays short. It simulates a fixed source machine halting after t steps on input of length n in O((t+1)(n+t+1)^2) time and O((t+1)(n+t+1)) space. For n = O(t) this is cubic time, though the constants are astronomically large. Lean 4 proofs establish universality, halting equivalence, these bounds, and recovery of the source machine's final configuration.
No takes yet. Share an insight, caveat, or question.
Michael Schroeder (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: