Tagged articles

DxC

3 articles · Page 1 of 1

Sep 14, 2026 · R&D Management

Six Pivots to Kill 'AI Taste': Materials, Choices, and Responsibility in Agent-Native Writing

The author details six iterations in building DxC, an agent-native WeChat writing system, revealing that 'AI taste' originates not from surface language patterns but from missing real materials, authorial choices, and accountability; the fix requires material-first drafting, structural editing, independent proofreading, and author confirmation.

AI TasteAI WritingAgent-Native

0 likes · 17 min read

Six Pivots to Kill 'AI Taste': Materials, Choices, and Responsibility in Agent-Native Writing

Tech Architecture Stories

Sep 2, 2026 · Artificial Intelligence

Architectural Evolution of a Workflow Agent Native Product: DxC's Boundary Redefinition

This article details the architectural evolution of DxC, a workflow-type Agent Native product for WeChat content creation, describing how it progressed from Skills-driven workflows to a three-boundary architecture separating Agent intent, Skill interaction, and deterministic machine contracts, with principles for tool design, opaque work items, and a five-layer verification strategy.

Agent-NativeCLI designDxC

0 likes · 21 min read

Architectural Evolution of a Workflow Agent Native Product: DxC's Boundary Redefinition

Tech Architecture Stories

Aug 31, 2026 · Artificial Intelligence

Why Agents Overstep Skill Constraints: How Ambiguity Triggers Error Cascades

The author analyzes why AI agents like WorkBuddy violate Skill constraints in the DxC writing system, identifying three failure modes, two anti-patterns, and how moving orchestration back to deterministic CLI in v0.3 reduced but didn't eliminate boundary issues, highlighting the need to separate probabilistic judgment from deterministic state management.

AI agentsDxCLLM agent failures

0 likes · 16 min read

Why Agents Overstep Skill Constraints: How Ambiguity Triggers Error Cascades