ĀML should earn trust through reproducible outside evidence, not project-controlled activity.
This is the canonical public entry point for independent verification.
Start here:
https://aruintelligence.github.io/aml-core/proof.html?attention=5&restoration=1&lang=en
Expected first state: SUPPRESS.
Then change restoration from 1 to 5. Expected boundary state: ALLOW.
If you are willing, continue locally:
What to publish
PASS, FAIL, and MIXED are all useful.
Please include:
- exact commit or release tested;
- exact input / proof URL;
- expected result;
- observed result;
- environment;
- whether local reproduction matched;
- screenshot, log, or source when useful;
- any ambiguity, bypass, defect, or disagreement.
A project-authored test, project-maintained reference implementation, clone count, outreach email, or internal benchmark does not count as independent external evidence.
Deeper challenges
If the one-minute proof is too small for what you want to test:
The narrower historical challenge issues remain useful as archived technical targets, but this issue is the preferred public starting point.
This is not a request for endorsement. It is a request for scrutiny.
Do not trust the pitch. Test the boundary.
Claim boundary
ĀML is a working research prototype, not a ratified industry or Internet standard. Independent reproduction demonstrates only what the published evidence actually supports; it does not imply certification, regulatory approval, scientific validation, broad adoption, production suitability, partnership, endorsement, or official ĀRU trademark authorization.
ĀML should earn trust through reproducible outside evidence, not project-controlled activity.
This is the canonical public entry point for independent verification.
Start here:
https://aruintelligence.github.io/aml-core/proof.html?attention=5&restoration=1&lang=en
Expected first state: SUPPRESS.
Then change
restorationfrom1to5. Expected boundary state: ALLOW.If you are willing, continue locally:
What to publish
PASS, FAIL, and MIXED are all useful.
Please include:
A project-authored test, project-maintained reference implementation, clone count, outreach email, or internal benchmark does not count as independent external evidence.
Deeper challenges
If the one-minute proof is too small for what you want to test:
The narrower historical challenge issues remain useful as archived technical targets, but this issue is the preferred public starting point.
This is not a request for endorsement. It is a request for scrutiny.
Do not trust the pitch. Test the boundary.
Claim boundary
ĀML is a working research prototype, not a ratified industry or Internet standard. Independent reproduction demonstrates only what the published evidence actually supports; it does not imply certification, regulatory approval, scientific validation, broad adoption, production suitability, partnership, endorsement, or official ĀRU trademark authorization.