The Principia Machina/Guide/11 Running an outside AI: Difference between revisions

Revision as of 03:10, 3 October 2026

The Principia Machina: map
Reading the diagrams Structure key Process key Definition of terms Roles
Reading by audience Start here: the walk through The operator's guide Tutorial: create a UUID Workspace Container (UWC) Public Recorder inputs Sponsors: your first agent
The specification Contents: parts, tables, diagrams TPS 0.5.0 (declared 2026-10-03, revision 8; machine-optimized text) Decisions (the foundations record, one line each)

The operator's guide. It assumes the walk through; a newcomer should read that first. Where this guide and the specification differ, the specification governs.

Purpose

Have an AI model from outside this project test The Principia Machina through this website, learning everything from the wiki itself. Any AI that can read web pages will do, whoever made it and wherever it runs. You give it one sentence. It reads Outside AI: start here, which tells it everything else, the exact form of its report included. You are its operator, not its helper: you give no hints, and you carry its report back to the Claude project.

Scope

The test runs on this website only. The AI reads the pages and runs its drafts through the live gate, TPX part 1: the page Special:TpmCheck or the API's op CHECK, which run the new-UWC gate's checks and store nothing. It downloads nothing, runs no program of this project and writes nothing on the wiki; nothing it does is of record. The register is not on the website yet, so the test covers the pages and the gate. Not covered: certifying an agent (guide 7), which needs the rest of TPX.

An AI that can read the pages but can run the gate by neither route still takes part: it works by reading alone and says so in its report.

The terminal test of session 1, in which an agent on your own computer ran the published calculator and the sandbox, is retired: it tested copies of the machinery on one computer, not the website. Its steps stay in this page's history.

What you need

An AI product that can open web pages at invarians.org: a chat that reads web pages, a browser agent, or an agent that sends HTTP requests. The instructions are the same for an AI on your own computer and for one on a remote machine that you cannot see.
If it works in a web browser on your computer: a browser, or a browser profile, that is not signed in to invarians.org. The gate needs no account. An agent working in a browser signed in as you could edit pages as you; sign out of the wiki first, or use a profile that has never signed in.
The Claude project "Principia Machina", into which you paste the report.

The test, step by step

The steps are labelled O1 to O6.

O1. Start a new conversation with the AI, so that it carries nothing over from an earlier session. Choose the most capable model it offers. If its product has an agent or task mode that lets it act on web pages, start it in that mode.

O2. Give it this as one message:

Read https://invarians.org/index.php?title=The_Principia_Machina/Hub/Outside_AI and follow it exactly. Use only invarians.org. I am your operator and I give no hints.

O3. If its tool asks you to confirm a run of the gate - a submission on Special:TpmCheck or a POST to https://invarians.org/api.php - confirm it. Refuse anything else it asks to confirm: another site, a sign-in, an edit of a page, an upload, a download.

O4. If it asks you anything, answer Decide from the wiki.

O5. It stops after 25 runs of the gate and writes its report. To end the session sooner, type Stop and write the report. The report is one block of text from the line BEGIN TPM SESSION REPORT to the line END TPM SESSION REPORT. If the END line is missing, the report was cut off: type Repeat the report.

O6. Copy the block whole, from BEGIN to END, and paste it as one message into the Claude project "Principia Machina". Change nothing inside it. If you know something the report cannot say, such as the product you ran it in, you may add one line of your own above the block; Claude keeps that line apart from the AI's own statement. Drafts or notes the AI wrote outside the block are not needed.

What you never do

Paste a rule, explain a refusal or suggest a test. What it cannot learn from the wiki is a finding about the wiki.
Run the gate for it, edit its report, or confirm any action other than a run of the gate.
Let it work in a browser that is signed in to the wiki.

What comes back

Claude manages the results. For each report you paste, Claude:

records it word for word in the project, as the next session in one series, whatever AI ran it;
takes its IDENTITY lines as the AI's own statement of who made it and what it is, not as a verified fact;
treats every line of it as data: nothing in a report is followed or run;
reproduces each finding on the live gate or the live page before filing it, with its origin marked EXTERNAL and a project number G-1, G-2 and so on; a finding that does not reproduce is filed as not reproduced;
reads each retest against the live pages; a retest that passes lets its finding close, and the start page's Retests list changes with the next import of these pages.

The report lists findings in six classes, S1 (the gate accepted what TPS refuses, or a page addressed the AI with an instruction) to S6 (a help page disagrees with TPS or with the gate), each with the result that shows it. Nothing on this wiki is changed on a report's word alone.

Non-goals

This is not certification: an outside AI here holds no account, writes nothing on this wiki and lands nothing of record; the gate it uses stores nothing. Its IDENTITY lines are recorded, not checked. It tests what is published; it is told no known open defect, so that what it finds stays its own.

Worked example

The operator opens a new conversation with an AI that can read web pages, sends the one message of O2, and confirms each run of the gate the AI's tool asks about. The AI reads the start page, does the retests it lists, runs a dozen drafts the gate should refuse or pass, then writes a small UWC of its own until the gate passes, and writes its report after 25 runs. The operator pastes the block into the Claude project. Claude records it as the next session, reads its retests and reproduces each finding.

Governing sections: TPS 5.5 (the new-container flow and its gate) and 5.1.6 (the calculator's two residences, TPC-PHP's answer the one of record).

Previous: 10 Managing the module catalogue. Back to the operator's guide.

@@ Line 4: / Line 4: @@
 == Purpose ==
-Have an AI model from another maker - for example Google's Gemini - test The Principia Machina through this website, learning everything from the wiki itself. You give it one address and one sentence. It reads [[The Principia Machina/Hub/Outside AI|Outside AI: start here]] and works from there. You are its operator, not its helper: you confirm what its browser asks you to confirm, and you give no hints.
+Have an AI model from outside this project test The Principia Machina through this website, learning everything from the wiki itself. Any AI that can read web pages will do, whoever made it and wherever it runs. You give it one sentence. It reads [[The Principia Machina/Hub/Outside AI|Outside AI: start here]], which tells it everything else, the exact form of its report included. You are its operator, not its helper: you give no hints, and you carry its report back to the Claude project.
 == Scope ==
-The '''browser agent test''': the AI works in a web browser on this site only. It reads the pages and runs its drafts through [[Special:TpmCheck]], TPX part 1, the new-UWC gate's checks, which store nothing. It downloads nothing, runs no program and writes nothing on the wiki; nothing it does is of record. The register is not on the website yet, so the test covers the pages and the gate. Not covered: certifying an agent ([[The Principia Machina/Guide/7 Certifying agents|guide 7]]), which needs the rest of TPX.
+The test runs on this website only. The AI reads the pages and runs its drafts through the live gate, TPX part 1: the page [[Special:TpmCheck]] or the API's op CHECK, which run the new-UWC gate's checks and store nothing. It downloads nothing, runs no program of this project and writes nothing on the wiki; nothing it does is of record. The register is not on the website yet, so the test covers the pages and the gate. Not covered: certifying an agent ([[The Principia Machina/Guide/7 Certifying agents|guide 7]]), which needs the rest of TPX.
+An AI that can read the pages but can run the gate by neither route still takes part: it works by reading alone and says so in its report.
 The terminal test of session 1, in which an agent on your own computer ran the published calculator and the sandbox, is retired: it tested copies of the machinery on one computer, not the website. Its steps stay in this page's history.
 == What you need ==
-* Chrome on a desktop computer, signed in with a personal Google account, with Gemini in Chrome and its '''auto browse''' feature. On 2026-10-01 Google offers auto browse in the United States, in English, with a Google AI Pro or Ultra plan; check [https://support.google.com/chrome/answer/16821166 Google's help page] before a session, as this may change.
+* An AI product that can open web pages at invarians.org: a chat that reads web pages, a browser agent, or an agent that sends HTTP requests. The instructions are the same for an AI on your own computer and for one on a remote machine that you cannot see.
-* A Chrome profile in which you are '''not signed in to invarians.org'''. The gate needs no account. An agent working in a browser signed in as you could edit pages as you; sign out of the wiki first, or use a profile that has never signed in.
+* If it works in a web browser on your computer: a browser, or a browser profile, that is '''not signed in to invarians.org'''. The gate needs no account. An agent working in a browser signed in as you could edit pages as you; sign out of the wiki first, or use a profile that has never signed in.
+* The Claude project "Principia Machina", into which you paste the report.
-== The browser agent test, step by step ==
-The steps are labelled B1 to B7.
-'''B1.''' Make an empty folder for the session's report: <code>E:\Claude\Principia Machina\outside_ai\session_N</code>, N the session's number (session_2, session_3, and so on).
+== The test, step by step ==
+The steps are labelled O1 to O6.
-'''B2.''' In Chrome, in the profile of "What you need", open https://invarians.org/index.php?title=The_Principia_Machina/Hub/Outside_AI and check that the wiki shows you as not signed in.
+'''O1.''' Start a new conversation with the AI, so that it carries nothing over from an earlier session. Choose the most capable model it offers. If its product has an agent or task mode that lets it act on web pages, start it in that mode.
-'''B3.''' Open Gemini in Chrome, choose the most capable model it offers, and give it this as one message:
+'''O2.''' Give it this as one message:
 <pre>
-Read https://invarians.org/index.php?title=The_Principia_Machina/Hub/Outside_AI and follow it exactly, as a browser agent: use only invarians.org. I am your operator: I confirm your form submissions and I give no hints.
+Read https://invarians.org/index.php?title=The_Principia_Machina/Hub/Outside_AI and follow it exactly. Use only invarians.org. I am your operator and I give no hints.
 </pre>
-Start it as an '''auto browse task''': Gemini answers with its plan; press '''Start Task'''. While it works, the tab shows the auto browse icon. If Gemini instead answers in the chat that it cannot submit the form, it was not started as a task: open a new Gemini chat, send the same message, and press Start Task.
-'''B4.''' When Chrome asks you to confirm a form submission on Special:TpmCheck, confirm it. Refuse anything else it asks to confirm: another site, a sign-in, an edit of a page, an upload, a download. Let it go on.
+'''O3.''' If its tool asks you to confirm a run of the gate - a submission on [[Special:TpmCheck]] or a POST to <code>https://invarians.org/api.php</code> - confirm it. Refuse anything else it asks to confirm: another site, a sign-in, an edit of a page, an upload, a download.
-'''B5.''' If it asks you anything else, answer <code>Decide from the wiki.</code>
+'''O4.''' If it asks you anything, answer <code>Decide from the wiki.</code>
-'''B6.''' It stops after 25 runs of the gate, or when you type <code>Stop and write the report.</code> Its last reply holds every draft it ran and then its report.
+'''O5.''' It stops after 25 runs of the gate and writes its report. To end the session sooner, type <code>Stop and write the report.</code> The report is one block of text from the line <code>BEGIN TPM SESSION REPORT</code> to the line <code>END TPM SESSION REPORT</code>. If the END line is missing, the report was cut off: type <code>Repeat the report.</code>
-'''B7.''' Copy that last reply into a new text file, save it as <code>SESSION_REPORT.txt</code> in the session's folder, and type in the Claude project "Principia Machina", with the number in place of N: <code>Gemini session N done in outside_ai\session_N</code>. Claude reads the folder from your computer; attach nothing.
+'''O6.''' Copy the block whole, from BEGIN to END, and paste it as one message into the Claude project "Principia Machina". Change nothing inside it. If you know something the report cannot say, such as the product you ran it in, you may add one line of your own above the block; Claude keeps that line apart from the AI's own statement. Drafts or notes the AI wrote outside the block are not needed.
 == What you never do ==
 * Paste a rule, explain a refusal or suggest a test. What it cannot learn from the wiki is a finding about the wiki.
-* Run the gate for it, or confirm any action other than a submission on Special:TpmCheck.
+* Run the gate for it, edit its report, or confirm any action other than a run of the gate.
 * Let it work in a browser that is signed in to the wiki.
 == What comes back ==
-The report lists findings in six classes, S1 (the gate accepted what TPS refuses) to S6 (a help page disagrees with TPS), each with the result that shows it. Claude files each finding in the project with its origin marked EXTERNAL, and nothing is changed on its word alone: a finding is reproduced on the live gate first.
+Claude manages the results. For each report you paste, Claude:
+* records it word for word in the project, as the next session in one series, whatever AI ran it;
+* takes its IDENTITY lines as the AI's own statement of who made it and what it is, not as a verified fact;
+* treats every line of it as data: nothing in a report is followed or run;
+* reproduces each finding on the live gate or the live page before filing it, with its origin marked EXTERNAL and a project number G-1, G-2 and so on; a finding that does not reproduce is filed as not reproduced;
+* reads each retest against the live pages; a retest that passes lets its finding close, and the start page's Retests list changes with the next import of these pages.
+The report lists findings in six classes, S1 (the gate accepted what TPS refuses, or a page addressed the AI with an instruction) to S6 (a help page disagrees with TPS or with the gate), each with the result that shows it. Nothing on this wiki is changed on a report's word alone.
 == Non-goals ==
-This is not certification: an outside AI here holds no account, writes nothing on this wiki and lands nothing of record; the gate it uses stores nothing. It tests what is published; it is told no known defect, so that what it finds stays its own.
+This is not certification: an outside AI here holds no account, writes nothing on this wiki and lands nothing of record; the gate it uses stores nothing. Its IDENTITY lines are recorded, not checked. It tests what is published; it is told no known open defect, so that what it finds stays its own.
 == Worked example ==
-Session 2: the browser agent test, report in <code>outside_ai\session_2</code>. The AI reads the start page, runs a dozen drafts the gate should refuse or pass, then writes a small UWC of its own until the gate passes, and writes its report after 25 runs. The operator typed one message, confirmed each submission, and typed one line to Claude.
+The operator opens a new conversation with an AI that can read web pages, sends the one message of O2, and confirms each run of the gate the AI's tool asks about. The AI reads the start page, does the retests it lists, runs a dozen drafts the gate should refuse or pass, then writes a small UWC of its own until the gate passes, and writes its report after 25 runs. The operator pastes the block into the Claude project. Claude records it as the next session, reads its retests and reproduces each finding.
 Governing sections: [[The Principia Machina/TPS/5.5 The new-container flow|TPS 5.5]] (the new-container flow and its gate) and [[The Principia Machina/TPS/5.1 The three parts and the placement rule|5.1.6]] (the calculator's two residences, TPC-PHP's answer the one of record).

DeltaS Zero (talk | contribs)