Be kind.
For AI readers: Machine-readable summary · llms.txt
Why?
Why be kind, rather than right? Why kindness, before truth, before authenticity, before freedom? And why take any of this seriously?
This work exists for one reason: to answer the question Why?
It starts with the only thing you can know for certain — one record exists.
That is the first principle. That is the axiom.
What is a record? A distinction that holds — this, not that. A footprint in sand is a record. So is a fossil. So is any bit of information. So are these words.
You are reading this sentence. That is a record. Try to deny it. The denial is a record.
From that first principle — one record exists — the rest follows by logic, step by step. There is one shared reality. The now is shared. Awareness is shared. The one being harmed is the one doing the harming, experienced from a different window. To harm the other is the same as the right hand hitting the left — it is the structure hurting itself. Therefore:
Don’t be a cunt. Be kind.
Nothing here compels you. You can be cruel. The logic forces nothing. It only shows you what is left. What is left is the price, and reality audits.
If you read one thing, read The Illusion of the Other. There is no physics in it. If it lands, nothing further is required.
The physics fell out of the same record. It was never the point. It gives the ethic its weight.
The whole of the derivation is on this site, free, forever. It starts here.
The front door, then the whole building. Nineteen rooms — start anywhere.
Curious? Start at the front door.
Sceptical? Start at the method, then the kill switches.
A physicist? Start at Ø Physics, then run the arithmetic at Ø Confirm the Math.
Nothing here asks to be believed. Every claim carries the condition under which it dies, and two of them have already died.
The 420 Code is a museum built to answer one question: why be kind? The answer is derived from reality, not commanded. Every claim hangs on the wall with the condition that would take it down. Everything starts with one first principle.
Never met an argument like this? Start with Three Ways of Being Sure — how the 420 Code knows anything at all. No equations. No prerequisites.
whybekind.global →The short answer.
aiethics.global →The ethics, applied to AI.
The heart of it. Five books, five ways into the same building — one of them opens for you.
You do not enter a building through the roof. You enter through a door. Each of these five opens into the same entrance hall.
The Illusion of the Other — the gentle door. The line you have drawn between yourself and every person you have ever met, erased. If you read one thing, read this.
Being After Religion — the front door. What produces the violence is not the people. It is the design.
Antichristos — the sacred door. Love your neighbour. Do not judge. Be kind. Is the teaching true? Not because a god said so, but because the structure of reality says so.
The Relationship Corridor: What Is Love — the personal door. Love is not a feeling. It is a sustained choice to keep the corridor wide.
The Interior — the operational door. AI alignment is a physics problem.
The catalogue — the philosophical structure. One axiom, run across five scales of question, from why there is something rather than nothing to how we live together, and beyond.
Four books, each installing the axiom once and running it across a different scale of question — and the first one written, Ø Predictions, superseded by the Frozen Predictions.
Dissolutions takes it to twelve problems philosophy has asked longest — why there is something rather than nothing, mind and body, free will, the gap between is and ought. Resolutions to thirteen more the tradition could not close. Applications to twelve architectures of how we live together — law, economics, medicine. Horizons to thirteen domains beyond — beauty, music, the sacred. And Predictions to physics — now superseded by the Frozen Predictions.
Rosin and the Records — the whole argument pressed into one voice, and the thirty years it came from, raw.
Rosin. The extract. Press heat into flower and what drips out is the thing itself — nothing added, nothing lost. Five books in five voices — prose, conversation, metaphor, nursery rhyme, proofs — each walking the whole argument in one compression. Prose and Proofs are out now.
The Records. The unprocessed output of a thirty-year project — mistakes, breakthroughs, dead ends. It began not in a library, but in a wound. Not polished, not organised, not safe: the truth before it learned to dress itself.
For the people who would rather watch than read.
Short films, one claim each, in G’s voice. The drawing draws itself, and the words write themselves as he says them. The films add his voice and the drawing, not a new claim.
They come in chains, each an argument in order, and every line a film says is on its page as text — so a chain reads whole with nothing playing.
One line. Everything else follows — and you proved its first principle by reading this sentence.
1:1 + 1×ε @ AS [+α | −α]
You are reading this sentence. That is a record. Try to deny it, and the denial is another. One record exists — and four conditions follow, not as assumptions but as properties the reading already has: Symmetry. Break. Record. Constraint.
Don’t take my word for it. AJ Greyling bought the books and didn’t believe a word of it. A harsh sceptic, he built his own machine to find out: a program that takes the published formulas and one measured number, recomputes every constant from scratch, and checks its working against my published script.
They match to machine precision. His computer, his code — and it runs offline, in one command, on yours. Go and try to break it. His own words on what that proves: “It does not prove the theory is true. Reproducing a derivation is not the same as nature agreeing with it.” That is what an independent check is for.
The same verifier keeps a plain-language companion to the frozen predictions at lucid.rodeo/prereg, and to the body of work at lucid.rodeo/the420code.
And the source checks itself. Every change re-derives every result from α and fails loudly if any drifts past its published tolerance.
One measured input. Zero free parameters. Everything else derived, unless stated otherwise.
Every theory of physics takes measured inputs. The Standard Model takes at least nineteen. This takes one: α, the fine-structure constant, about 1/137.
From it: how much heavier the proton is than the electron, and the neutron than the proton; how strong gravity is; what the universe is made of, how old it is and how fast it is growing; and the acceleration below which galaxies stop following Newton — each set against its measurement. Two registered predictions have fired. They stay in the table, shown, never repaired.
The proton. The count meets the measurement at ten parts in a trillion. When the laboratories tighten their bar to 3.4 parts in a trillion with the central value unchanged — expected within three to five years — the row dies at three sigma. Frozen as mp-me-closed.
Gravity. The realised value sits about three hundred and sixty parts per million below the adopted average. If the adopted value settles upward as the laboratories’ systematics close, the prediction of G dies outright, with no fallback. KS-CCC.3.
The expansion rate. The closure took a side in the Hubble tension. If the rate, as the data settle it, lands outside 64.4–70.8 either way, the closure dies. KS-ASM.1.
The visible share. Nearest to firing. If the visible fraction settles above 4.81% at better than half a per cent precision — expected within about five years — the row dies. KS-41.1.
The one interior. If the fine-structure constant is shown to vary by place or age, at the discovery standard, by two independent methods, the strongest claim in the work dies. KS-GLASS.3.
Every one of these was frozen with a date, in public, before the measurement. The full register, with every document and script: Ø Frozen Predictions.
Every result against its measurement — and what else is derived →
Written down and time-stamped before the measurement that will test them — so no one, G included, can quietly change them after.
A scoreboard, not a showcase. The live predictions and the two that fired stand side by side: a fired switch is shown, never repaired, never deleted. The same freezes are read independently, in plain language, by the verifier who anchors them.
One number goes in. Every headline result comes out. Run it here, in your browser, and see for yourself.
One measured input. Zero free parameters. The code below reproduces every headline result from a single number — the fine-structure constant α ≈ 1/137. Copy it. Paste it into any Python environment. Run it. If the numbers match, the arithmetic holds. Whether each formula is forced is argued in its paper, and every step carries its switch. If the numbers don’t match, the argument has a problem and you’ve found it.
from math import pi, log, exp, cos, tan, sqrt, asinh
# ─── ONE MEASURED INPUT ───
alpha = 1 / 137.035999177 # fine-structure constant (CODATA 2022)
# ─── PHYSICAL CONSTANTS (CODATA 2022) ───
hbar = 1.054571817e-34 # reduced Planck constant (J·s)
c = 299792458 # speed of light (m/s)
m_e = 9.1093837139e-31 # electron mass (kg)
G_measured = 6.67430e-11 # gravitational constant (N·m²/kg²)
mp_me = 1836.152673426 # m_p/m_e (CODATA 2022)
mn_me = 1838.68366200 # m_n/m_e (CODATA 2022)
KM_PER_MPC = 3.0857e19
GYR = 365.25 * 86400 * 1e9
# ─── CLAIM 1: Proton-electron mass ratio — the series closed (AP49) ───
# AP30 stopped at α² with a coefficient of sixteen it declared owed under KS-30.3.
# AP49 The Hold paid it: r = 16α/1836 and the series closes as a chain of holders.
scaffold = 21**2 * 4 + 21 * 3 + 3**2 # = 1836
maintenance = alpha * 21 * (1 - 1 / (84 * pi)) # dynamic term
hold_share = 16 * alpha / 1836 # r, the share repaired each order
correction = 21 * alpha * hold_share / (1 - hold_share) # all orders
ratio_pred = scaffold + maintenance + correction
print("PROTON-ELECTRON MASS RATIO (AP49 — the series closed)")
print(f" Predicted: {ratio_pred:.8f}")
print(f" Measured: {mp_me:.8f}")
print(f" Error: {abs(ratio_pred - mp_me) / mp_me * 1e9:.3f} ppb")
print()
# ─── CLAIM 2: Gravitational constant — realised (AP44) ───
# G = α²¹ × (1 + 1/π) / (1 + α) × ℏc/mₑ² the arena's held ε discounts the coupling once
alpha_G = alpha**21 * (1 + 1 / pi)
G_first = alpha_G * hbar * c / m_e**2 # AP28’s first reading, +0.69 % — withdrawn
G_real = G_first / (1 + alpha) # AP44: −0.036 %
print("GRAVITATIONAL CONSTANT (AP44)")
print(f" Realised: {G_real:.4e} N·m²/kg² ({(G_real - G_measured) / G_measured * 100:+.3f}%)")
print(f" Measured: {G_measured:.4e} N·m²/kg²")
print()
# ─── CLAIM 3: Neutron-proton mass difference — realised (AP47) ───
# Δ_real = Δ_bare / (1 + α²/8π); the bare row FIRED on 2026-08-02 at 7.24σ and stays beside it
delta_bare = 3 * (1 - 1 / (2 * pi)) + alpha * (1 + 1 / (2 * pi))
delta_real = delta_bare / (1 + alpha**2 / (8 * pi))
delta_meas = mn_me - mp_me # 2.530988574, bar 7.4e-7 (0.29 ppm)
print("NEUTRON-PROTON MASS DIFFERENCE (AP47 realised; AP30 bare row fired)")
print(f" Realised: {delta_real:.11f} mₑ ({(delta_real - delta_meas) / 7.4e-7:+.3f} σ)")
print(f" Bare row: {delta_bare:.8f} mₑ ({(delta_bare - delta_meas) / 7.4e-7:+.2f} σ — FIRED)")
print(f" Measured: {delta_meas:.9f} mₑ")
print()
# ─── CLAIM 4: Dark sector partition (AP42) ───
f_DM_dark = (6/21) * (1 - exp(-21/6))
f_DE_dark = 1 - f_DM_dark
f_vis = 1 / 21
f_DM = f_DM_dark * 20 / 21
f_DE = f_DE_dark * 20 / 21
print("DARK SECTOR PARTITION (AP42)")
print(f" Dark energy: {f_DE * 100:.2f}% (observed: 68.89%)")
print(f" Dark matter: {f_DM * 100:.2f}% (observed: 26.07%)")
print(f" Visible matter: {f_vis * 100:.2f}% (observed: ~4.885 +/- 0.05%)")
print()
# ─── CLAIM 5: The age of everything — one cycle (AP46) ───
# t = (21/18) · α⁻¹⁸ · τ_C one escape through eighteen doors on the electron's tick, at the duty 18/21
tau_C = hbar / (m_e * c**2)
t_cycle = (21/18) * alpha**-18 * tau_C
print("AGE OF THE UNIVERSE (AP46)")
print(f" Predicted: {t_cycle / GYR:.3f} Gyr (one lane-time wide: ±0.66)")
print(f" Measured: 13.787 ± 0.020 Gyr (Planck 2018)")
print()
# ─── CLAIM 6: The expansion rate — the closure (AP48) ───
# H₀t₀ = (2/(3√Ω_Λ))·arcsinh√(Ω_Λ/Ω_m) at the work's own partition; the age is the cycle
omega_L, omega_m = f_DE, f_DM + f_vis
H0t0 = (2 / (3 * sqrt(omega_L))) * asinh(sqrt(omega_L / omega_m))
H0 = H0t0 / t_cycle * KM_PER_MPC
print("EXPANSION RATE (AP48)")
print(f" Pure number H₀t₀: {H0t0:.3f}")
print(f" Predicted: {H0:.2f} km/s/Mpc (window 64.4–70.8)")
print(f" Measured: 67.4 ± 0.5 km/s/Mpc (Planck 2018)")
print()
# ─── CLAIM 7: MOND acceleration scale at the work's rate (AP18) ───
# a₀ = C_S²·c·H₀/(2π) where C_S² = 2·ln(sec(½) + tan(½)) ≈ 1.0445 — not the fine-structure constant
CS2 = 2 * log(1 / cos(0.5) + tan(0.5))
a0_pred = CS2 * c * (H0 / KM_PER_MPC) / (2 * pi)
a0_meas, a0_unc = 1.20e-10, sqrt(0.02e-10**2 + 0.24e-10**2) # random ⊕ systematic
print("MOND ACCELERATION SCALE (AP18, at the work's H₀)")
print(f" C_S²: {CS2:.10f}")
print(f" Predicted: {a0_pred:.4e} m/s²")
print(f" Measured: {a0_meas:.4e} ± {a0_unc:.2e} m/s² ({(a0_meas - a0_pred) / a0_unc:.2f} σ low)")
print(f" The floor inverted, 74.3 ± 1.2 (KS-45.1): {(74.3 - H0) / 1.2:.1f} σ from the work's rate — FIRED 2026-09-03")
print()
print("─" * 50)
print("One input (α). Zero free parameters. Two corpses shown.")
print("Copy this code. Run it. Confirm it yourself.")
Runs entirely in your browser — offline. Nothing installed, no network, nothing sent anywhere. It re-derives the same checks as verify.py (the script the badge above runs on every change), at the same double precision — the two outputs match line for line.
Fifty-four proofs, bound in eight notebooks — from the premise to its consequences, each with a summary in plain words.
The Premise. Spacetime. Quantum Mechanics. Forces & Constants. Particles & Matter. Cosmology. The Operator Interface. Consequences. Every Artist’s Proof, bound by subject, each notebook with its summary.
Six hundred and four kill switches. Two have fired. They stay on the wall.
Every claim carries a stated condition under which it dies. The switches are not hidden. They are published — all six hundred and four, with the two that have fired, on one page.
Kill Switch Registry — v5.41, 15 November 2026 PDF
Each with its status, the kind of test that would fire it, and the paper it comes from.
The words the work is built from, defined one by one.
The complete vocabulary, notation and imagery of the body of work. Every term that does structural work — across the eight Notebooks, the five Ø Models, the five Five Doors books, and the Rosin Ø series — drawn into one canonical reference.
17 thematic families. 425 entries, 391 index entries. Edition 1.5, 11 November 2026: three new headwords — now, time and awareness — and every grade and line read as the registry reads at v5.39: where a rule is meant, it is a stated rule, and no grade rests on anyone’s ruling. No sense moves. From the axiom and its four conditions through to the falsification apparatus, the applied frame, the relational corridor, religion and hermeneutics, AI alignment, the cost family and the hourglass.
Structural Glossary — edition 1.5 PDF
Every term on one page — its definition, the paper that installs it, and the grade it carries.
The work is free. The cost is not.
Support the work now has its own room at Studio G, the production house and custodian of the 420 Code: what is still planned, where the support goes, and where the work stands, in 21 languages.
Support the work →One record exists.
Be kind is a derivation.
The awareness in me is the awareness in you.