Are LLMs killing microservices? What once required fleets of specialized ML models and fragile orchestration can now run inside a single Python agent. Software complexity is rising. Architectural comp…
資訊系統設計
無論是哪種資訊系統、不管是 architecture 還是 infrastructure 都歡迎討論
Follow from your instance
@information-system-design@baudrate.tw
Search for this handle on your own server to follow from there.
This week's engineering radar: 📖 READ: File Mover Express: AWS Releases Open-Source Tool for S3 File Transfers [News] Renato Losio highlights AWS's new open-source tool for transferring large local fi…
Raw intelligence is no longer the only bottleneck. Speed and cost matter just as much. As Chip Huyen explains on the #InfoQ Trends Report podcast, growing demand for cheap, low-latency AI is reshaping…
AI agents are powerful. But what happens when they browse, execute code, and call tools with limited human oversight? Docker highlights 5 security risks that emerge when agents operate autonomously - …
Your app’s performance, release speed, and stability at scale depend on what sits behind it. This #InfoQ article explores the architecture behind modern applications, why it matters, and the challenge…
What belongs in the shared AI platform? The InfoQ AI Engineering Program with Hien Luu examines this in a live online cohort. Review the syllabus: https://certification.qconferences.com/ai/ai-engineer…
Anthropic calls it the Supervision Paradox: Overuse AI ➔ Lose coding skills ➔ Can’t verify AI. The catch? Supervising AI requires the exact technical depth that over-relying on it destroys. If junior …
How does LinkedIn connect user activity to backend performance? LinkedIn has built an OpenTelemetry-based telemetry platform that traces interactions across its iOS, Android, and web apps through the …
Better #AIagents don't need more context. They need the right context. Reliable agentic AI starts with #ContextEngineering, not prompt engineering. The key? ➤ Give agents relevant information. ➤ Balan…
Java keeps moving. What’s new this week? 🔗 From TornadoVM 8.0’s GA release to Hibernate’s latest beta releases, catch up on the Java ecosystem’s latest updates, tools, and releases 👉 https://www.infoq…
Measure flow, not only output. The InfoQ High-Performing Teams Program with Olimpiu Pop examines this in a live cohort. Review the syllabus: https://certification.qconferences.com/high-performing-team…
Want to reduce LLM hallucinations? Start with your documentation, not better prompts. Suhail Patel, Staff Engineer at Monzo, explains how well-structured documentation can become tools and skills that…
Lectures don't make great Software Architects. Real-world collaboration, scars from production failures, and continuous adaptation do. As AI takes on more architectural work, knowing what to delegate …
Are junior developers becoming obsolete? As AI agents take over routine coding, engineering teams are quietly freezing junior hiring. On the AI, ML & Data Engineering Trends Report, Daniel Dominguez a…
If AI writes the code, who trains the next generation of engineers? The routine tasks AI handles today - writing boilerplate, fixing bugs, and building basic features - were never just tasks. They wer…
Where does context lose its freshness? The InfoQ AI Engineering Program with Adi Polak examines this in a live cohort. Review the syllabus: https://certification.qconferences.com/ai/ai-engineering/nov…
800+ MCP servers. 5,000+ tools. One gateway. Uber has built an MCP Gateway to help AI agents discover and invoke internal services at scale. The architecture brings together automated API discovery, a…
Check out what's new in the #Spring community 👉 https://www.infoq.com/news/2026/10/spring-news-roundup-oct19-2026/ First release candidates of: Spring Boot, Spring Framework, Spring Data, Spring Cloud…
AWS has recently released File Mover Express - an #opensource application for transferring large files between local systems and Amazon S3. Available for macOS, Windows, and Linux, it supports paralle…
When AI writes 90% of your codebase, traditional code review breaks down. Anthropic’s Claude Code team shared what their process looks like: • 50+ automated review rounds overnight on a single PR • AI…