Agent systems are advancing quickly across domains, but their evaluation remains fragmented. Most benchmarks rely on fixed, LLM-centric harnesses that require heavy integration, create test-production mismatch, and limit fair comparison across diverse agent designs. The root problem is the lack of an open, agent-agnostic assessment interface. We advocate Agentified Agent Assessment (AAA), where evaluation is performed by judge agents and all participants interact through standardized protocols: A2A for task management and MCP for tool access. Conventional benchmarking defines two separate interfaces, one for the benchmark and one for the agent, while AAA only needs one; this yields a generic, unified framework that separates assessment logic from agent implementation and enables reproducible, interoperable, and multi-agent evaluation. We further introduce AgentBeats as a concrete realization of AAA: we identify five practical operation modes that make standardized assessment compatible with real-world constraints on openness, privacy, and reproducibility. To evaluate our design at scale, we conduct two studies: a five-month open competition that drew 298 judge agents across 12 categories together with 467 subject agents from independent participants, showing that AAA applies across a heterogeneous range of benchmarks; and a case study on coding agents that confirms agentified evaluation preserves fidelity with the public record while surfacing previously missing head-to-head results, yielding research insights about agent design. Combining a community-scale field study and a controlled coding case study, we verify that AAA delivers coverage, practicality, and fidelity across heterogeneous scenarios at scale. Together, AAA and AgentBeats offer a clear path toward open, standardized, and reproducible agent assessment.
This JAMA Forum discusses state-level momentum and progress of local programs to address social determinants of health in the context of cuts made by the Trump administration to federal programs.
Policymakers have contemplated multiple Social Security reform plans as the combined OASDI Trust Fund approaches depletion in 2034, when tax revenues will fund only 81% of scheduled benefits. The authors argue that as the depletion date approaches without meaningful reform efforts, general revenue financing for Social Security - at least for a period of time - becomes increasingly likely. The magnitude of required adjustments far exceeds historical precedent: restoring 75-year solvency would require an immediate and permanent 22% benefit cut or 29% payroll tax increase (3.65% points). Absent reform, benefits would be cut by 19% upon depletion in 2034, rising to 28% by 2099. Since 1977, the maximum benefit decrease averaged across all beneficiaries has been just 0.7%, while the payroll tax rate has remained fixed at 12.4% since 1990. Using novel stochastic simulations, the authors demonstrate that even reforms eliminating the 75-year actuarial deficit would leave the system vulnerable to repeated depletion threats. Without building a substantial buffer stock - requiring even larger changes than typically proposed - Social Security faces a 40-50% probability of Trust Fund exhaustion within decades under plausible economic scenarios. The authors present implementation options for using general revenues, including temporary versus permanent financing mechanisms and redirecting existing revenue streams. They provide a conceptual analysis of the efficiency and equity consequences, weighing intergenerational fairness and political sustainability against otherwise unavoidable benefit disruptions.
Health policy is undergoing a period of rapid transformation, driven by major health reforms, changing market structures, pharmaceutical policy developments, technological innovation, and evolving models of care delivery. These changes have profound implications for health care access, affordability, quality, and health system sustainability. In this editorial, we identify 10 priority areas for health services and policy research that we believe will shape policy debates and evidence needs in the coming years. These priorities are organized across 5 domains: health care access, coverage, quality and delivery system reform; corporatization and financialization of health care systems; pharmaceutical policy and health services research; international comparisons and cross-national learning; and innovation, technological advancements, and resiliency of health systems. Specific areas we discuss include the effects of the H.R.1 on Medicaid coverage and economic security; the future of Medicare Advantage; the impact of immigration policies on access; the consequences of increasing corporatization of health care; reforms to pharmaceutical pricing; the role of international comparative research; and the implications of precision nutrition, climate-related disruptions, and artificial intelligence for health policy and health system performance. Across these topics, we emphasize the need for rigorous, policy-relevant research that moves beyond descriptive analyses to evaluate mechanisms, implementation, and real-world impacts.
BACKGROUND:Engaging in moderate-to-vigorous physical activity (MVPA) in early childhood can have both immediate and long-term health benefits. Strategies for consistently supporting this are currently unknown, largely due to the vast number of potentially interrelated and dynamic contributing factors that may also be heterogenous across children. METHODS:We developed an agent-based model (ABM) that represents children ages 3-9 who can engage in MVPA in a variety of settings in which they spend time. Our model incorporates key theoretical constructs identified in the literature and is grounded in high-quality empirical evidence, primarily participant data from a cohort-based randomized controlled trial with extensive longitudinal accelerometry measurement. We assess the ABM's ability to reproduce patterns of MVPA observed in the cohort. RESULTS:We uncovered a specific model representation of key pathways and settings involved in MVPA for this age group that can closely reproduce real-world MVPA across multiple a priori assessment metrics. This specification provides new insights into modeled contributors to MVPA. Within our model, the most important within-model pathway driving MVPA for girls is the quality of the built environment, while, for boys, it is the social environment; given the relative availability of the two, this might explain observed differences in MVPA in real-world settings. CONCLUSIONS:In addition to immediate insights, this article provides proof of principle for a powerful tool that can further explore the etiology of childhood MVPA and inform practices and policies to positively affect early childhood PA, setting the stage for lifelong health.
