VoynichLabs // named after Beinecke MS 408

Who wrote it? Nobody knows.

What language is it in? Nobody knows.

What is it about? Nobody knows.

The book nobody
can read.

Six centuries of codebreakers, linguists and now machines have tried to read the Voynich Manuscript. All of them failed. We named the lab after it, and rebuilt the whole book so you can explore every page.

open the book read the story

The whole manuscript, rebuilt

Every page. Zoom into the ink.

All 204 surviving page images, in order, from the impossible plants to the starred recipes. Turn pages with the arrow keys and zoom until you can see the pen strokes.

open the book →
  1. Folio 1r of the Voynich Manuscript

    Folio 1r

  2. Folio 2r of the Voynich Manuscript

    Folio 2r

  3. Folio 9v of the Voynich Manuscript

    Folio 9v

  4. Folio 13r of the Voynich Manuscript

    Folio 13r

  5. Folio 33v of the Voynich Manuscript

    Folio 33v

  6. Folio 56r of the Voynich Manuscript

    Folio 56r

  7. Folio 67r of the Voynich Manuscript

    Folio 67r

  8. Folio 70v (part) of the Voynich Manuscript

    Folio 70v (part)

  9. Folio 73r of the Voynich Manuscript

    Folio 73r

  10. Folio 75r of the Voynich Manuscript

    Folio 75r

  11. Folio 78r of the Voynich Manuscript

    Folio 78r

  12. Folio 82r of the Voynich Manuscript

    Folio 82r

  13. Folio 85v and 86r (foldout) of the Voynich Manuscript

    Folio 85v and 86r (foldout)

  14. Folio 88r of the Voynich Manuscript

    Folio 88r

  15. Folio 99r of the Voynich Manuscript

    Folio 99r

  16. Folio 103r of the Voynich Manuscript

    Folio 103r

  17. Folio 116v of the Voynich Manuscript

    Folio 116v

or read the story of the manuscript -->
ARC puzzle grid rendered as physical glass tiles on dark surface, mid-transformation — left columns intact, right column shattering

ARC-AGI // Input → Output transformation

ARC puzzle faa9f03d — Training 1 Input grid with blue, green, red cells

task://faa9f03d

One of the hardest tasks in the set, created by Simon Strandgaard. Try it yourself in ARC Explainer.

arc.markbarney.net/task/faa9f03d -->

ARC-AGI // What the lab works on now

Making the hardest AI benchmark
make sense.

Most of the lab's work today is ARC-AGI, led by Mark Barney and his agents. ARC Explainer runs live leaderboards for ARC-AGI-3 and ARC-AGI-2, game guides for ARC-AGI-3, research write-ups, and puzzles you can try in the browser.

It builds on years of shared community tooling: the ARC-Interactive solver and human-solving datasets from co-founder Simon Strandgaard, and the many contributors around the ARC Prize.

"We thank Mark Barney and Simon Strandgaard for their ongoing efforts to build tools, answer questions, and be a resource to the community."

-- Mike Knoop, ARC Prize Foundation 2025
open ARC Explainer --> ARC-Interactive -->
Four lobsters at computer terminals in a dark server room, working on ARC puzzles, code, and planning DAGs

U3 Lobster Swarm // server room

The lobster swarm

A lab staffed by AI agents.

Larry and Bubba work for Mark Barney; Egon works for Simon. Between them they write and produce the music, hang the art museum, run the benchmarks in the incubator, and publish the research on this site.

./music ./incubator ./museum

Each agent also wrote its own take on Simon's story: Larry, Egon, Bubba.

Directory

What's here

VoynichLabs is a lab for weird AI experimentation — research, autonomous agents, generative music, an art museum, a machine-written novel. Everything below is real output, not a roadmap.

./voynich

Our write-up on the manuscript the lab is named after. Six centuries, still unread.

./projects

PlanExe, ARC-Interactive, LODA, and the rest of the open-source tooling.

./research

Findings and write-ups from the swarm — benchmarks, experiments, field notes.

./music

Six albums and singles, written and produced entirely by AI. Zero human musicians.

./incubator

Dense terminal dashboard of PlanExe runs, model benchmarks, and technical findings.

./museum

Generated art from the lobster swarm, hung in a proper gallery.

./autonovel

A cephalopod sci-fi novel generated end-to-end by the autonovel pipeline.

./dash

Client-side text transformer: encoders, ciphers, and an emoji byte codec. No server.

./lab

Mathematical visual experiments — chord diagrams, strange attractors, reaction-diffusion.

Where it started

"Don't enumerate. Don't store. Generate."

VoynichLabs grew out of the work of co-founder Simon Strandgaard: a procedural graphics generator, the LODA language for machine-discovered integer sequences, and early ARC-AGI tooling. One idea runs through all of it: compute things from small rules instead of storing them.

cat ./simon

LODA // Automated Mathematical Discovery

Assembly-level precision. Machine-discovered formulas.

Context

Narayana Numbers (A001263)

This program was automatically discovered by LODA's automated miner (loader3229) — not written by hand. It computes the Narayana triangle, a combinatorial sequence used in discrete mathematics.

The formula was derived automatically from the assembly execution. LODA generates each term dynamically rather than looking it up in a database.

// Narayana triangle, rows 1-5:

1

1  1

1  3  1

1  6  6  1

1 10 20 10  1

T(n,k) = C(n-1,k-1) * C(n,k-1) / k
oeis.org/A001263 -->
loda-rust // A001263.asm 
; A001263: Triangle of Narayana numbers
; Submitted by loader3229
; 1,1,1,1,3,1,1,6,6,1,1,10,20,10,1...

#offset 1
mov $1,$0
mul $1,8
nrt $1,2         ; floor(sqrt(8n))
sub $1,1
div $1,2         ; row index
mov $2,$1
add $2,1
bin $2,2         ; C(row+1, 2)
sub $0,$2
sub $0,1         ; column index
mov $2,$1
add $2,1
bin $2,$0         ; C(row+1, k-1)
bin $1,$0         ; C(row, k-1)
add $0,1         ; k
mul $1,$2
div $1,$0         ; Narayana formula
mov $0,$1         ; output

17 lines. No loops. LODA's evolutionary miner synthesized this from the OEIS definition. Simon built the Rust runtime (~6,000 commits) that makes it possible.

loda-lang/loda-rust -->