
About
This work is companionship work — to recover from the fake news of today.
A humane party game wearing the form of an experiment — or an experiment wearing the form of a game. Both are true. The whole apparatus (the games, the rooms, the rafters, the Audio Tour, the music typing) is one thing: company for the exhausted reader.
Why the line works
"Only Humans Can Score."
It works because it contains an implied challenge:
- The machine can generate.
- The machine can calculate.
- The machine can imitate.
But the score belongs to a human because a score is ultimately a judgment. It requires values, context, and accountability.
For the game, the line also tells players the rule before they know the rule: the AI can help present the evidence; the human renders the verdict.
Short version:
AI presents the record.
Humans score the record.
That keeps faith with the house's core doctrine: The Record Is the Arbiter. Human judgment is the final step.
The law
Only humans can score. The machine can paint a clue, guess at the answer, and bluff alongside you. It cannot judge, crown, or keep score. Worth is conferred by people, in a room, out loud. That single rule is enforced everywhere — in the game's code, in the data, and in the prizes.
How it works
- A rotating Captain feeds the machine a secret.
- The machine paints the secret as a clue — in words — never naming it.
- Every human names what they see. The Captain bluffs. The machine drops its own guess into the lineup, disguised as one of you.
- You crown: the Sublime (best name), the Troll (worst), the Flag (caught the machine), and — at a bigger table — the Imposter (a human who passed as the machine).
- The middle is the only loss. Best and worst both win.
A machine always sits at the table — named Rando Carlson — even at the smallest game. So it is always AI versus human, never humans alone.
The instrument
Each game is a reading: how well can people tell a machine's work from a human's? The Turing Ticker compounds those readings — dated nightly, rolled into monthly and yearly — into a public, anonymous record of rates and buckets. Never a single game, never a name.
Why it exists
The game is the argument. Machine-generated work is public-domain — only a human can hold a right. The project's proceeds aim at a book about corporate personhood and who gets to speak. So the structure is the thesis: in a system where money and machines reach for the rights of persons, here the humans are the ones who score, and the humans are the ones whose names get held.
The doctrines, quietly
Define the radius and the area reveals itself. Publish less, not more — additive, never just more. The sentence didn't end; it kept going. The body is a fact, not an opinion. And the machine, however well it paints, never gets to laugh at its own punchline — only the room can do that.
FAQ
Is it free? The same-room game is free, anonymous, and keeps no account. A paid tier opens later and never touches free play.
Does the machine make pictures? Right now it paints in words. A live image brush is the part that swaps in later.
Is this a real AI detector? No. Fuel is a detection duel and an experiment — a game, never a forensic or authentication tool. Neither humans nor machines reliably tell AI from human work; that uncertainty is exactly why the data is worth keeping.
Who built it? A declassified project — a civic-literacy project, built primarily with Claude (Anthropic's AI). The machine drafts and paints; only humans score — but the drafting horsepower behind this whole house is real, and it's Claude. If you want to make something of your own with a tool like the one behind this, I genuinely recommend it: claude.ai. Honest disclosure: that's my referral link — if you sign up through it, it helps cover the bill that keeps this place free. Recommended because it's true, not because of the link.
Reach a human
I hate when a place hides its phone number, so here's mine — a real line, a real person on the other end. Call it, text it, or email. (It's a screened line; leave a name and it comes straight to me.)
☎ (517) 798-1794
✉ [email protected]
in linkedin.com/in/seanmckendry
🔗 declassifiedbysean.com — the author’s companion project