default author photo

Tatsuya Shimomoto

Independent

SCHOLARLY PAPERS

3

DOWNLOADS

35

TOTAL CITATIONS

0

Ideas:
“  I'm currently working on design principles for AI coding agents — how they should accumulate and curate knowledge (the Agent Knowledge Cycle), attribute the ideas they reuse (Agent Attribution Practice), and act with contemplative restraint — and the broader question underneath all of it: what "authorship" and "self" mean once ideas diffuse primarily through LLMs rather than citations. Research practice note: My work routinely involves AI assistance (primarily Claude, Anthropic) for structural critique, prose refinement, and code drafting. I retain accountability for all substantive claims, framing, and final judgments.  ”

Scholarly Papers (3)

1.
Downloads 17 (1,518,554)

Abstract:

Loading...

AI accountability, autonomous agents, attribution gap, LLM workflow, phase separation, AI governance, moral crumple zone

Abstract:

Loading...

AI Accountability, Black Box, Scaffolding, Minimum Disclosure, Causal Traceability, Prohibition Hierarchy, AI Governance, Autonomous Agents

3.

Harness Alignment and Harness Drift: Why Intent, Unlike Correctness, Resists Automation

Number of pages: 12 Posted: 29 Jun 2026
Tatsuya Shimomoto
Independent
Downloads 8

Abstract:

Loading...

Harness Alignment, Harness Drift, Intent Alignment, LLM Agents, Agent Harness, Human-In-Theloop, Software Evolution, Automation Complacency