More Podcasts by InfoQ episodes
Published 4 Mar 2026
Duration: 00:52:05
Generative AI is transforming software architecture, requiring a shift from incremental tool adoption to systemic thinking and governance to manage unpredictability and mitigate risks.
This conversation explores why generative AI is not just another automation layer but a shift into autonomy. The key idea is that we cannot retrofit A...
The podcast explores the complexities of integrating generative AI into modern software architecture, highlighting the shift from incremental tool adoption to systemic redesign. It emphasizes the risks of hallucination and drift when embedding AI into procedural systems, arguing that retrofitting AI without rethinking governance and architecture can lead to unpredictable outcomes. The discussion underscores the need to move away from rigid control mechanisms toward boundary management, where clear goals, constraints, and seven defined "seams" (interfaces and governance points) are established to contain AI agents and ensure alignment with organizational objectives.
Key challenges include managing emergent behaviors in multi-agent systems and adapting guardrails as systems scale. The podcast stresses the importance of redefining architectural principles to accommodate AI autonomy, balancing innovation with robust governance frameworks. It advocates for proactive strategies to mitigate risks while enabling systems that are both scalable and adaptive, ensuring that AI integration enhances rather than undermines stability and functionality in complex software environments.
11 May 2026 From Java EE to Quarkus and LLMs: Adam Biens Playbook for Boring, FutureProof Systems
A veteran Java developer highlights preferences for minimal dependencies, critiques enterprise Java's complexity, advocates for Quarkus' lightweight design over Spring, emphasizes Java EE standards, legacy BCE patterns, and Java's stability in AI/LLM integration and cloud challenges.
4 May 2026 Roq: Leveraging Quarkus to Build Static Sites at the Speed of Go
Java's resurgence is fueled by performance gains, modern frameworks like Quarkus, and native compilation, exemplified by Rooka lightweight static site generator leveraging Quarkus for dynamic rendering, Markdown content, and streamlined workflows, with future AI integration and open-source advancements.
27 Apr 2026 A Java Performance Quest: Taming Unsafe Code, Embracing Idiomatic Style & Debugging the Linux Kernel
Challenges in deploying AI from proof-of-concept to production highlight engineering gaps, with QuestDB's three-tier storage (real-time, query, archival) and Java/C++/Rust optimizations enabling scalable time-series data handling, though facing Java warmup, code generation, and hardware integration hurdles.
20 Apr 2026 Engineering Stable, Secure and Scalable Platforms: A Conversation with Matthew Liste
Systems engineering and software development's evolution emphasizes hands-on learning, mentorship, and intuitive experience, while addressing AI's impact on apprenticeships, balancing abstraction with deep system understanding, managing risks in high-stakes sectors, navigating innovation-stability trade-offs, scaling complex systems, evolving engineer roles, customer feedback loops, fostering continuous learning and collaboration, and prioritizing craftsmanship, systemic thinking, and the synergy between technical precision and practicality.
13 Apr 2026 How SBOMs and Engineering Discipline Can Help You Avoid Trivys Compromise
Strengthening mobile app security beyond minimal standards, leveraging Software Bill of Materials (SBOM) to address supply chain risks under legislative mandates like U.S. Executive Orders and the EU's Cyber Resilience Act, and utilizing tools such as Cyclone DX and SPDX for dependency tracking, compliance, and mitigating supply chain attacks through improved tooling and practices like OIDC authentication.