Jul 2026· ACM SIGOPS Operating Systems Review· Vol 60, pp. 21 - 29· 1 citation· ⚡ 1 influential· 56 references
TL;DR
It is shown that solving systems problems such as resource scheduling and code optimization requires reasoning that is fundamentally as hard as any reasoning problem a human might have to solve, which makes systems problems a promising target for AGI research, and that both the AI and systems community should investigate this area more.
Abstract
Computer systems are increasingly using AI techniques, as evidenced by increasing numbers of publications and several reported deployments in production systems. While early work in this area focuses on relatively small ML models and statistical techniques that replace heuristics, the field is experiencing a paradigm shift towards using agentic techniques to automatically discover new systems algorithms and designs, driven by a rapid increase in model capabilities. We argue that these recent advances in AI capabilities – specifically, models that reason about the world and perform complex sequences of actions through agentic interactions – not only suggest that the OS community should employ AI more widely in its research, but also that it should look at AI for Systems through a new lens. We show that solving systems problems such as resource scheduling and code optimization requires reasoning that is fundamentally as hard as any reasoning problem a human might have to solve. By establishing a strict success criterion for what it means to ''fully solve'' these tasks across open-ended, real-world edge cases, we demonstrate that these problems are ''AGIcomplete''. We argue that this makes systems problems a promising target for AGI research, and that both the AI and systems community should investigate this area more.
Overall, it is found that the use of coding agents in scientific computing holds great promise for accelerating scientific research and increasing the reliability of critical systems, but that outstanding concerns remain.
Jeremiah H. Li, Alex Rubinsteyn, Sergey Feldman et al.· bioRxiv· 0 citations
This article examines the emerging paradigm of agentic AI for scientific discovery, traces the conceptual shift from tools to agents, lays out a six-stage workflow spanning literature synthesis to manuscript generation, and reviews practical systems in chemistry, equation discovery, materials science, and general machine learning research.
Alexander Taktakidze· Longevity Horizon· 0 citations
Large language models have evolved from curiosity to co-pilot in under four years. With the emergence of agentic AI systems that reason, plan, and execute multi-step tasks autonomously, HPC centers face a new category of user need: researchers expect these tools to be supported, not just permitted. This paper offers a practitioner’s perspective from Purdue’s Rosen Center for Advanced Computing (RCAC), where we have begun deploying system-wide configurations, custom MCP servers, and user guidance for agentic tools. As a demonstration, every word of this manuscript was produced through an agent-first workflow: over one hundred commits of iterative collaboration between human authors and AI agents, documented in a public GitHub repository. We argue that proactive engagement, not prohibition, is the path forward for facilitators who wish to remain relevant in the agentic era.
Geoffery Lentner, Ashish· Practice and Experience in A...· 0 citations
AI-driven autonomous research (AR) systems are becoming increasingly effective across a broad range of tasks. Their performance, however, is still evaluated primarily by the quality of the final outcome. In this paper, we argue that the efficiency of the solution-search process is an equally important but often overlooked dimension of performance. A strong AR system should not only produce high-quality results, but also reach them with as small a budget as possible. Search efficiency will become increasingly important as AR expands from domains with inexpensive verification, such as mathematics and coding, to real-world scientific settings in which solution evaluation may require costly physical experiments. To capture this dimension, we propose evaluating AR systems using the area under the curve (AUC) of the Pareto frontier, alongside final outcome quality. We compare several families of search algorithms, including hill climbing, beam search, tree search, and evolutionary search, across twelve systems-optimization tasks. We find that no single search structure is consistently the most efficient. We also show that search efficiency and final outcome quality are distinct performance dimensions: a method that eventually achieves the best result may nevertheless improve slowly and consume substantially more evaluation budget before reaching that result. Because the most effective search policy is generally unknown in advance, we introduce an adaptive procedure called fluid search, which uses a portfolio bandit to dynamically allocate a fixed evaluation budget across a forest of search processes. Across the evaluated tasks, fluid search achieves the highest overall search efficiency, closely matching the performance of a per-task oracle that is given the best search structure for each task in advance.
Artificial intelligence has evolved from specialized tools to comprehensive, goal-oriented systems that can learn, adapt, and soon even reason. New models that can write code, produce art, forecast market movements, and support decisions in all spheres of society are being developed every week. But this book covers more than just technology. Power exists. Ethics are a factor. It has to do with intelligence, the nature of future work, and people in general. It is up to the pioneers, leaders, and curious thinkers who must now lead us through the most disruptive time in our history.
Dr. Jörg Storm
is an international keynote speaker, podcast host, and lecturer at several academic institutions. He holds a Ph.D. in Economics and is recognized for turning technological complexity into crisp, board-level decisions and measurable business outcomes.
This position paper outlines a common abstraction layer that can substantially lower the effort for designing agents with constrained creativity, and demonstrates early promise of this paradigm for two SysOps use cases: Root Cause Analysis and Network Configuration Generation.
†. SayanSinha, V. Harsh, Yajie Zhou et al.· 0 citations
