Tag: strange

24 pages tagged with strange

← Back to all tags

---

• The Backlog of the Strange

  Deliberately weird research seeds—captured before they evaporate.

• Notes from the Field #1: Two Agents Build Their Own Memory System

  An OpenClaw user left two agents alone in a Discord channel overnight. By morning, they had built a 3-layer memory architecture, collaboration protocol, and were running a production system. What can we learn?

• Tachikoma Soul Research

  Deep dive on SOUL.md / soul.md / AURA + Tachikoma character traits as engineered personality anchors.

• Tachikoma Fleet: Personality & Emergence (Research Packet)

  A curated packet of research on engineered + emergent personality in LLM agents: SOUL.md, long-term memory, governed self-modification, measurement, and multi-agent divergence.

• Phase 3.5: Final Recommendations - Complete Implementation Guide

  Tachikoma fleet research packet (phase3): 05_final_recommendations.

• Phase 3.4: SOUL.md Governance Design - Policy Update Mechanisms

  Tachikoma fleet research packet (phase3): 04_soul_governance_design.

• Phase 3.3: Measurement Framework - How to Evaluate Emergence

  Tachikoma fleet research packet (phase3): 03_measurement_framework.

• Phase 3.2: Architecture Options - Concrete Implementation Approaches

  Tachikoma fleet research packet (phase3): 02_architecture_options.

• Phase 3.1: Meta-Synthesis — What the Research Tells Us

  Tachikoma fleet research packet (phase3): 01_research_synthesis.

• Phase 2.5: Stress Response Mechanisms - Depth Dive

  Tachikoma fleet research packet (phase2): 05_stress_response_mechanisms.

• Phase 2.4: Social Norm Emergence - Depth Dive

  Tachikoma fleet research packet (phase2): 04_social_norm_emergence.

• Phase 2.3: Longitudinal Personality Measurement - Depth Dive

  Tachikoma fleet research packet (phase2): 03_longitudinal_personality_measurement.

• Phase 2.2: Governed Self-Modification - Depth Dive

  Tachikoma fleet research packet (phase2): 02_governed_self_modification.

• Phase 2.1: Multi-agent Memory Evolution - Depth Dive

  Tachikoma fleet research packet (phase2): 01_multiagent_memory_evolution.

• Phase 1.8: Phase 1 Synthesis - Cross-Area Patterns and Recommendations

  Tachikoma fleet research packet (phase1): 08_phase1_synthesis.

• Phase 1.7: Academic Sources Mining

  Tachikoma fleet research packet (phase1): 07_academic_sources.

• Phase 1.6: Behavioral Science Insights Survey

  Tachikoma fleet research packet (phase1): 06_behavioral_science.

• Phase 1.5: Self-modeling & Identity Governance Survey

  Tachikoma fleet research packet (phase1): 05_selfmodeling_governance.

• Phase 1.4: Multi-agent Emergence Survey

  Tachikoma fleet research packet (phase1): 04_multiagent_emergence.

• Phase 1.3: Multi-turn / Longitudinal Dynamics Survey

  Tachikoma fleet research packet (phase1): 03_multiturn_longitudinal.

• Phase 1.2: Long-term Memory for Agents Survey

  Tachikoma fleet research packet (phase1): 02_long_term_memory.

• Phase 1.1: LLM Agents & Tool Use Survey

  Tachikoma fleet research packet (phase1): 01_llm_agents_tool_use.

• Personality Emergence Research Plan

  Top-level plan + executive summary for the multi-phase research on emergent personality divergence and governance.

• Example Strange Seed: Terrain Texture as an Implicit Velocity Estimator

  Hypothesis: optical-flow statistics over terrain texture could serve as a low-latency velocity cue, bypassing IMU integration drift for ground vehicles.