Obelisk General Examination¶

The Obelisk General Examination (OGE) is the proposed capstone acceptance process for broad human competence. It borrows one key idea from All Souls: let expert humans judge generated work under difficult, open-ended conditions, then interrogate the candidate’s reasoning in a viva. It extends that model beyond essays.

Components¶

1. General Judgment I - three unseen essays across civilization, ethics, science, institutions, technology, art, life and uncertainty.
2. General Judgment II - mixed source/data/object packet; synthesize and make a decision for an educated non-specialist audience.
3. Specialist Depth - chosen trunk or craft; serious technical/scholarly depth.
4. Systems & Institutions - a consequential case requiring technical, human-world and normative reasoning.
5. Practical Stations - rotating live tasks: health-information navigation, financial/contract reasoning, repair/safety triage, visual/data critique, measurement, conversation/negotiation.
6. Learning Sprint - an unfamiliar domain disclosed only at the start; learn enough to solve, explain and teach, then return for a delayed retest.
7. AI Stewardship Trial - delegate a consequential task under permissions while hidden failures test provenance, evals, monitoring and rollback.
8. Viva Voce - 45–90 minutes with multiple examiners who challenge the candidate’s own work, uncertainty and revisions.

Human acceptance criterion¶

The strongest version should combine analytic rubrics with blind whole-dossier ranking by independent experts. Examiners answer a consequential question:

Would I trust this person to enter an unfamiliar, ambiguous, high-stakes problem; recognize what they do not know; learn quickly; work with experts and AI; act responsibly; and improve the situation?

That judgment should never stand alone, but it captures something a checklist cannot.

See data/obelisk/general-examination.json for the machine-readable design.