Agent frameworks
4 mentions across all digests
Software architectures for orchestrating autonomous AI agents, with research showing single-agent LLMs can outperform multi-agent systems on multi-hop reasoning under equal token budgets.
To Throw a Stone with Six Birds: On Agents and Agenthood
arXiv paper rigorously defines what constitutes agents and agency from first principles—foundational work as agentic AI systems become mainstream.
Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgets
Single-agent LLMs beat multi-agent orchestration on multi-hop reasoning under equal token budgets, suggesting simpler agent architectures may be more computationally efficient than specialized multi-agent setups.
Let's Have a Conversation: Designing and Evaluating LLM Agents for Interactive Optimization
Research identifies structured dialogue patterns that improve LLM agent design and evaluation for interactive optimization, establishing reusable blueprints for autonomous development tools.
Glia: A Human-Inspired AI for Automated Systems Design and Optimization
Glia, a multi-agent LLM system, autonomously designs distributed systems algorithms that rival human-expert solutions—demonstrated via GPU cluster request routing and scheduling optimization.