Uploaded June 2026 | Updated September 2026, 2 weeks ago
Two of the fastest-growing open-source AI agent projects sit down to discuss what comes next for agent frameworks, memory, learning, and autonomous workflows.
In this conversation, leaders from OpenClaw and Nous Research share the stories behind their breakout projects, discuss the evolution of agent memory systems, dynamic skill creation, and long-term learning, and debate the tradeoffs between open-source agent frameworks and vertically integrated model-plus-agent systems. The discussion explores how agents can become more personalized over time, why memory architecture matters, and what separates great agent experiences from today’s generation of coding assistants.
The panel also covers the rise of agent-native interfaces, autonomous skill generation, community-driven development, and the future of open-source AI. Learn how two of the most influential agent projects are thinking about reliability, personalization, model training, and the next generation of agent infrastructure.
Key Takeaways
• OpenClaw and Hermes take different approaches to agent development, with OpenClaw emphasizing flexibility and ecosystem integrations while Hermes focuses on tightly integrating memory, learning, and model training.
• Long-term memory is becoming a foundational capability for AI agents, enabling systems to retain context, preferences, workflows, and learned behaviors across sessions.
• Dynamic skill creation allows agents to automatically learn from mistakes and create reusable workflows that improve future performance.
• Open-source agent frameworks can move faster and experiment more aggressively than enterprise products constrained by legal, compliance, and platform requirements.
• The best agent experiences increasingly come from actions agents take automatically rather than capabilities users must explicitly request.
#OpenClaw #NousResearch #HermesAgent #AIAgents #OpenSourceAI #AgentFrameworks #LLMEvals #AIEngineering #ArizeObserve
🔗 Try Arize AX & Phoenix OSS: arize.com
🔔 Subscribe for weekly content on LLMs, agents, and evaluation: youtube.com/@arizeai?sub_confirmation=1
Two of the fastest-growing open-source AI agent projects sit down to discuss what comes next for agent frameworks, memory, learning, and autonomous workflows.
In this conversation, leaders from OpenClaw and Nous Research share the stories behind their breakout projects, discuss the evolution of agent memory systems, dynamic skill creation, and long-term learning, and debate the tradeoffs between open-source agent frameworks and vertically integrated model-plus-agent systems. The discussion explores how agents can become more personalized over time, why memory architecture matters, and what separates great agent experiences from today’s generation of coding assistants.
The panel also covers the rise of agent-native interfaces, autonomous skill generation, community-driven development, and the future of open-source AI. Learn how two of the most influential agent projects are thinking about reliability, personalization, model training, and the next generation of agent infrastructure.
Key Takeaways
• OpenClaw and Hermes take different approaches to agent development, with OpenClaw emphasizing flexibility and ecosystem integrations while Hermes focuses on tightly integrating memory, learning, and model training.
• Long-term memory is becoming a foundational capability for AI agents, enabling systems to retain context, preferences, workflows, and learned behaviors across sessions.
• Dynamic skill creation allows agents to automatically learn from mistakes and create reusable workflows that improve future performance.
• Open-source agent frameworks can move faster and experiment more aggressively than enterprise products constrained by legal, compliance, and platform requirements.
• The best agent experiences increasingly come from actions agents take automatically rather than capabilities users must explicitly request.
#OpenClaw #NousResearch #HermesAgent #AIAgents #OpenSourceAI #AgentFrameworks #LLMEvals #AIEngineering #ArizeObserve
🔗 Try Arize AX & Phoenix OSS: arize.com
🔔 Subscribe for weekly content on LLMs, agents, and evaluation: youtube.com/@arizeai?sub_confirmation=1
