Socrates
Holds the question open and refuses premature certainty.
The first product inside 1o1.org
The intelligence here isn't artificial.It's yours.
Not a better chatbot. Not a better prompt. Socrates is a memory-bearing dialectic engine that helps an Actor surface the assumption they did not know they had.
The Problem
Prompt engineering improves the prompt. Socrates improves the inquiry loop.
RLHF improves average model behavior across many users. General chatbots optimize for answer completion or engagement. Socrates is pointed at a different unit of value: the Actor's movement from hidden assumption to clearer next question.
“Your AI asks really good questions to challenge the questioner, help them get to the truth of their situation, and find their own solution.”
Early tester feedback · counselor by profession
The Architecture
Three layers. One inquiry loop.
No single component carries the whole differentiation. The system becomes different because the components are pointed at the same object: not the answer, but the evolution of inquiry.
Layer 01
Elastic Dialogue
Sprint keeps the disciplined five-step spine. Chat lets the Actor pause, backtrack, ask why, follow a tangent, and return only when the conversation has earned synthesis.
Actor-ledAnti-scriptStill disciplined
Layer 02
Memory + Source Context
Actor memory, graph context, and uploaded documents enter the same inquiry field. The point is not generic retrieval; it is bringing external material into contact with internal patterns.
Actor graphDocumentsLongitudinal
Layer 03
Persona Lenses + Artifact
Socrates, Strategist, Mirror, and Challenger pressure the same question from different vectors. The session leaves behind an artifact, so the inquiry accumulates instead of evaporating.
4 lensesArtifactPortable insight
The Method
Two modes. One spine. Different depth.
Sprint is the structured artifact-producing mode. Chat is the elastic layer: open-ended enough to stay alive, disciplined enough not to collapse into generic conversation.
01
Ask
Bring a real question, wish, fear, or decision.
02
Clarify
Find what is actually being asked.
03
Pressure
Surface assumptions, tensions, and missing evidence.
04
Synthesize
Name the movement without closing too early.
05
Return
Let the artifact and memory shape the next session.
The Lenses
Four lenses. One Actor.
Personas are not characters for performance. They are cognitive lenses that pressure the Actor's question from different directions while staying grounded in the same memory and context.
The Strategist
Tests sequence, constraints, tradeoffs, and next action.
The Mirror
Reflects what the Actor’s own language reveals.
The Challenger
Attacks the position so the Actor can see what holds.
2Depth modes
4Persona lenses
20Documents per Actor
1Artifact per completed Sprint
If it’s real, you’ll feel it in one interaction.
Bring a real question. If the system works, you will not leave with more output. You will leave with a better question.