The field of AI and software engineering is rapidly evolving, with a significant focus on enhancing the infrastructure and frameworks that support AI agents and autonomous systems. A notable trend is the development of agent infrastructure, which aims to mediate and influence AI interactions within open-ended environments, ensuring accountability and shaping interactions to mitigate risks. This includes the creation of shared protocols and technical systems that facilitate the integration of AI agents into existing legal, economic, and digital service frameworks.

Another key area of advancement is the integration of Large Language Models (LLMs) into software systems, particularly in autonomous vehicles and digital markets. This integration is transforming capabilities in natural language understanding, decision-making, and autonomous task execution. However, challenges remain in ensuring these systems are secure, private, and interoperable, highlighting the need for comprehensive software reference architectures.

In the realm of software engineering, there is a growing emphasis on making research software FAIR (Findable, Accessible, Interoperable, and Reusable), with projects like SoFAIR aiming to improve the discoverability and reusability of open research software. Additionally, the field is seeing advancements in the evaluation of architectural patterns in machine learning systems, with frameworks being developed to quantitatively assess their impact on scalability and performance.

Security in software engineering is also receiving attention, with exploratory studies shedding light on how developers engineer security features and the challenges they face. This research is crucial for understanding the practical needs of developers and improving the security of software systems.

Microservices architecture continues to be a focal point, with studies exploring the evolution of cloud-native systems, the impact of code changes across microservices, and the detection of semantic dependencies. These efforts aim to enhance the scalability, resilience, and maintainability of microservices-based systems.

Finally, the importance of intentional data scaling in AI development is being recognized, with arguments for prioritizing tasks that benefit most from data scaling and considering the topology of data in shaping future compute paradigms.

Noteworthy Papers

• Infrastructure for AI Agents: Proposes the concept of agent infrastructure to mediate AI interactions, emphasizing accountability and interaction shaping.
• Beyond the Sum: Analyzes infrastructure requirements for AI agents in digital markets, highlighting barriers and potential for economic efficiency.
• Making Software FAIR: Introduces SoFAIR, a project aimed at improving the FAIRness of research software through machine-assisted workflows.
• Systems Engineering for Autonomous Vehicles: Advocates for the use of LLMs in AV development, offering a proof-of-concept for supervisory control.
• A quantitative framework for evaluating architectural patterns in ML systems: Presents a framework for assessing the impact of architectural patterns on ML system scalability and performance.
• An Exploratory Study on the Engineering of Security Features: Provides empirical insights into how developers engineer security features, validating common assumptions.
• Towards Change Impact Analysis in Microservices-based System Evolution: Proposes infrastructure for change impact analysis in microservices, aiming to mitigate ripple effects.
• Semantic Dependency in Microservice Architecture: Introduces the Semantic Dependency Matrix for detecting and managing semantic dependencies in microservices.
• A Functional Software Reference Architecture for LLM-Integrated Systems: Describes a preliminary functional reference architecture for LLM-integrated systems, addressing key architectural concerns.
• Network Centrality as a New Perspective on Microservice Architecture: Explores the use of centrality metrics for assessing MSA quality and detecting anti-patterns.
• Myriad People Open Source Software for New Media Arts: Presents a dataset of open source projects and contributors in new media art, highlighting the intersection of software and art.
• Not Every AI Problem is a Data Problem: Argues for intentional data scaling in AI development, considering the topology of data.
• Software Bills of Materials in Maven Central: Mines SBOMs from Maven Central, providing insights into SBOM publication practices.