TOOLS / WISDOM / IMPACT

Sharpen the Weapon
Being built

Sharpen the Weapon

Iron sharpeneth iron; so a man sharpeneth the countenance of his friend.

Proverbs 27:17, KJV

We're aiming to be the world's greatest AI agent for scripture. No team gets there alone, and no AI gets there by grading itself. It gets there because people who know the Word keep sharpening it.

That's you. You don't have to be a scholar. If you know scripture, you can help. If you also use AI tools well, you can help even more, as long as you check what the AI tells you against a source you've opened yourself.

Here's the deal. You fix it where you read it. A reviewer who didn't write it checks your source. If you're right, the card is corrected in public, with your name in the log. If two faithful readings genuinely disagree, we don't vote. We show both, each argued at its strongest.

Testers can already flag errors with in Bible Weapon. The full correction form, the reviewer step and the public ledger are planned. Badges are built into the next test build. Everything on this page marked "planned" is exactly that.

How it works

  1. Flag it.

    Tap on the verse or card that's wrong.

    In testing
  2. Fix it where you read it.

    The form already knows which verse or card you're on. It asks three things: what's wrong, what it should say, and your source. If you used an AI tool, say which one, and cite a source you opened yourself. Your flag stays on your device until you choose to send it.

    Being built
  3. We sort it.

    A typo, a fact or a matter of doctrine, each goes to the right lane.

    Planned
  4. Someone else reviews it.

    Reviewers earn the right, open the source themselves, and never review their own work.

    Planned
  5. It goes live, with credit.

    The corrected card ships with a public changelog line and your credit in the ledger, and it becomes a test case for the public accuracy score.

    Planned

Who can help

  • You know scripture. That's enough. Bring the verse, the context and your source.
  • You know scripture and AI. Use your tools to hunt errors fast, then prove each one from a source you opened. AI can find a lead, but only a source settles it.
  • Everyone else: read, learn and follow along. All of it is open to read.

We don't put doctrine to a vote.

Some questions faithful Christians answer differently. On those, a majority doesn't make anyone right. So we don't count hands. We show each position, argued at its strongest by people who hold it, with its sources, and we let you weigh them with your Bible open.

Evidence strength

Planned
Submitted
Sent in with at least one source you can open. Not yet verified.
Checked
At least three verifiers (two for a typo), never the person who sent it in, each opened a source, and no sourced objection is left.
Established
At least five verifiers, including people who usually disagree, two independent primary sources, 30 days with no challenge upheld, and a maintainer's sign-off.

On doctrine, a level means each position is stated fairly. It never means one side is right.

People supply the evidence. AI may sort it, but it never writes it.

Correction log

Every correction: the claim, the evidence and how strong it is.

This shelf is being stocked. Check back soon.

DDT · Digital danger tape

Before you go in

Everything on this side of the door is still being built, and we share it before it is finished, on purpose.

We are asking the community to help us get it right. Where something looks wrong, flag it. Every flag is weighed against the evidence, and the evidence decides how strong an entry is: Submitted, Checked or Established.

That is also how any AI we build will be kept honest. It will answer only from entries the community has checked, so the data we give it has to be right first.

Anything nobody has checked yet says Draft. Please read it as a draft.