◀ THE FOLD0ROOT.AI // WORLD II · GRIND · THE GRINDSTONE◆ .dlw.fold
THE FOLD / GRIND / THE GRINDSTONE / THE BUSY BEAVER

THE BUSY BEAVER

the longest-running halter — and the edge of the computable
1 WHAT IT IS · WHAT IT DOES · FACT OR FICTION
The busy beaver asks a deceptively simple question: among all n-state, 2-symbol Turing machines that halt when started on a blank tape, which runs the longest, and which prints the most 1s? Call those record values S(n) and Σ(n).

The shock is that these functions grow faster than any computable function — Σ(n) is a concrete, finite thing that no algorithm can compute. Already S(5) is 47,176,870 and Σ(6) exceeds 10↑↑15. For n=3 the champion prints Σ(3)=6 ones (the longest-running 3-state machine, a different one, takes S(3)=21 steps).

LIT verified live: the canonical 3-state champion, from a blank tape, halts in 14 steps having written exactly 6 ones; the 2-state champion halts in 6 steps with 4 ones (window.__busybeaver). FIG no framing; these are exact simulations of documented machines.
2 HOW IT WAS WEAVED · AI + HUMAN
David (human) seated this at the-grindstone — the machine that grinds hardest and longest, then stops. The busy beaver is the grindstone’s patron: the halter that works the most before falling silent. AVAN (AI) built the instrument: the Turing-machine simulator, the champion tables, the space-time diagram.

Credit as content: Tibor Radó posed it in On non-computable functions (1962); Shen Lin & Radó settled n=3 (1965); recent collaborative work settled S(5) (2024). The weave: David names the hardest worker; I run the exact champions to a halt, then show the horizon of uncomputability behind them.
3 ONE DIMENSION
The tape as a line of cells, the head reading and writing as it steps through the champion. A finite machine, a finite program, and yet the only way to learn how long it runs is to run it — there is no shortcut in general.
4 TWO DIMENSIONS · INTERACTIVE
Choose the 2-state or 3-state champion and step it, or run to the halt. Watch the tape fill and the head shuttle; the machine stops itself at the busy-beaver record.
5 THREE DIMENSIONS + AVAN’S INVERSE
The green forward object: the space-time diagram — each row a snapshot of the tape, stacked in time — the champion’s finite, halting trace laid out as terrain.
AVAN’s addition (the inverse-companion): the forward question is ‘find the machine that runs the longest and still halts.’ Its inverse is a wall: to know Σ(n) you must know which machines halt — and that is the halting problem, undecidable in general. So Σ is perfectly well-defined (there are only finitely many n-state machines) yet not computable; past a certain n, even ZFC set theory cannot prove its value. The inverse of ‘the maximum’ is ‘the unknowable’: a finite question whose answer no algorithm can produce, because the non-halters never announce themselves. Magenta is that horizon — the machines still running, that may halt in a step or never; green is the champion’s trace, the last thing computation can say before a silence you cannot predict. The busiest beaver marks exactly where knowing ends.
LIT Genuine busy-beaver champions (Rado 1962; Lin & Rado 1965 settled n=3). Verified live by exact Turing-machine simulation: the documented 3-state champion (A:0->1RB,1->1RH; B:0->0RC,1->1RB; C:0->1LC,1->1LA) halts from a blank tape in 14 steps writing 6 ones = Sigma(3); the 2-state champion halts in 6 steps / 4 ones = Sigma(2) (window.__busybeaver). S(3)=21 (max steps) is achieved by a different machine — reported honestly, not conflated with the max-ones champion.

FIG No framing: the interpreter and both champion tables run in-browser to a genuine halt. The AVAN inverse is honest and is a real theorem — Sigma(n) is finite and well-defined but not computable because deciding which machines halt is the halting problem; the magenta 'horizon' represents the undecidable non-halters, not a computed value.
◆ sealed .dlw.fold → folded to ROOT_0 · a sphere of THE GRINDSTONE · David Lee Wise (ROOT0), with AVAN