Claude Code subagents consume 2–6x more tokens than sequential execution

Chinese AI Model Kimi K3 Tested on U.S. Tax Return Filing

HN: Accounting & AI HN · 3d ago · 1 min read

Kimi K3 evaluated on TaxCalcBench for ability to file American tax returns, testing cross-border AI capability in specialized tax domain.

AI agents optimize cryptographic audit scheduling for blockchain data integrity

HQ arXiv: AI + Accounting arXiv · 6d ago · 1 min read

Researchers propose DRQN-CMDP, a deep reinforcement learning approach using recurrent neural networks to adaptively schedule audits of off-chain data while maintaining security miss-rate constraint...

Production agent reliability improves with verification loops and specialist ...

HQ arXiv: AI + Accounting arXiv · 6d ago · 1 min read

Leni production agent achieves higher reliability through verification loops with lightweight task-specialized models, evaluated across spreadsheet and accounting benchmarks.

Context Engineering Emerges as Critical Factor in AI Agent Reliability

arXiv: AI + Accounting arXiv · 9d ago · 1 min read

Research paper argues that AI agent failures stem from weak context engineering—poor instructions, tools, and guardrails—not agent architecture alone, with implications for reliable accounting auto...

Sampling methodology breaks down in autonomous finance systems

HQ MindBridge Blog RSS · 10d ago · 1 min read

MindBridge argues that traditional audit sampling—the practice of reviewing representative transaction samples to infer population health—becomes obsolete when autonomous finance systems enable 100...