The article argues that recent executive overreach and politicization have severely undermined established norms of U.S. civil-military relations. To restore professional standards, the incoming Congress must actively exercise its oversight powers, moving beyond mere criticism to demanding transparency in defense policy and personnel decisions. Key reforms include enforcing meritocratic promotion processes, requiring documented justification for senior officer removals, and preventing the partisan use of the military. These actions are crucial for bolstering public trust and ensuring the military remains a nonpartisan instrument of civilian control.
Simpler Is Better for Autograders: Toward Cost-Effective LLM Evaluations for Open-Ended Tasks
English Summary
This RAND report addresses the bottleneck of evaluating large language models (LLMs) in open-ended tasks, which is typically constrained by the high cost and slow speed of expert human grading. The analysis tested five autograding methods and found that the simple 'single rubric' approach consistently outperformed complex techniques like metaprompting or prompt optimization. This method achieves a statistically significant reduction in error while matching or exceeding the accuracy of nonexpert human graders, but at a fraction of the time and cost. Policymakers should adopt single-rubric autograders as the default, scalable solution to enable cost-effective and reliable LLM evaluation across diverse domains.
中文摘要
這份RAND報告探討了評估大型語言模型（LLMs）在開放式任務中的瓶頸問題，該瓶頸通常受限於專家人工評分的高成本和低效率。分析測試了五種自動評分方法，發現簡單的「單一評分標準」（single rubric）方法，持續優於諸如元提示（metaprompting）或提示優化（prompt optimization）等複雜技術。此方法在顯著降低錯誤率的同時，其準確性可與非專家人工評分相當甚至超越，但所需的時間和成本卻大大降低。政策制定者應將單一評分標準的自動評分器作為預設的、可擴展的解決方案，以實現跨多領域、具成本效益且可靠的LLM評估。
Related Entries
-
1.
-
2.
The article argues that Donald Trump's unique and prolonged lame-duck period, coupled with domestic political pressures, makes him dangerously prone to international instability. Having exhausted major diplomatic gambits, Trump is expected to channel his energy into high-stakes, unilateral actions, primarily targeting adversaries such as Iran and Cuba. These potential actions—ranging from fresh strikes in the Persian Gulf to attempts at regime change—carry a significant risk of destabilizing regional security and triggering a global economic crisis. Policymakers must anticipate this pivot to high-risk foreign policy to mitigate potential international flare-ups.
-
3.
The Blue Wave signals significant voter frustration with cost of living and inflation, giving Congress a mandate to prioritize affordability and cost-reduction measures. Economically, the rapid build-out of AI data centers is creating localized stress, driving up skilled labor wages and straining resources, while also facing constraints from power grids and potential local regulatory backlash. Policymakers must manage this 'sequencing issue'—addressing rising costs before full productivity gains from AI can materialize—and anticipate shifts in global data center infrastructure as the market matures.
-
4.
The built environment workforce, comprising 15.3 million workers across 95 occupations, is foundational to modern infrastructure and economic stability. While highly paid, the workforce suffers from significant demographic imbalances, being disproportionately white and male, and often lacking clear career progression pathways. To prevent project delays and cost overruns, policy must mandate that capital asset owners treat labor market forecasting as core to project planning. Strategic interventions are required to strengthen career pipelines through robust apprenticeships, improve retention, and actively recruit younger and more diverse workers.
-
5.
The U.S. is drastically reducing core funding to multilateral institutions, withdrawing from numerous bodies due to perceived misalignment with national interests. While multilateral channels offer critical advantages—such as pooling resources for global public goods and managing risk in fragile settings—the U.S.'s funding remains heavily 'earmarked,' which effectively 'bilateralizes' the system and undermines shared burden-sharing. This practice, coupled with deep cuts, diminishes the multilateral system's effectiveness and efficiency, particularly in humanitarian aid. Policy recommendations suggest the U.S. must adopt a systematic, published strategy for aid channel choice and shift from restrictive project-level earmarking to broader thematic funding to restore global cooperation.