Reddit Discussion: Real-World Architectures and Habits for AI Agents
Ambitious-Prompt-975 · reddit · 2026-08-06
A developer started a thread to break out of their bubble and understand the community's actual habits and tech stacks for AI agents. The discussion covers several core dimensions:
- Deployment & Invocation: Are agents running locally or hosted? How are they triggered?
- Frameworks: Are developers using SDKs or frameworks like LangChain? How complex is it to implement protocols like A2A or MCP?
- Memory & Interaction: How is persistent memory handled? Is the interaction via text, voice, or video avatars?
- Security & Multi-Agent: Do users care about security? Are agents communicating with each other?
The poster invited others to share whether they build agents for business or fun, and how their workflows have changed over the past year.
Related event: Developers Discuss Real-World AI Agent Frameworks(3 posts)→
More from coding & agent
- Organizing Multiple MCPs: A Lightweight Reliability Layer for Orchestration — FewScarcity6957 · 2026-08-06
- Microsoft Exec: Agentic Coding Is Reshaping Development, Will Expand to All Departments — clamanna · 2026-08-06
- The Next AI Scaling Law: Letting Models Live Longer to Accumulate Experience — imjustnewatai · 2026-08-06
- Useful Prompt: Let AI Summarize Context for a Seamless New Session — doodlestein · 2026-08-06
- Meta Launches Muse Code: Parallel Agents for Real-Time Game Generation — qinzytech · 2026-08-06
- Does AI Summarizing Execution Experience Count as Self-Improvement? — ___Patrice___ · 2026-08-06