We've announced our Spring '26 Camp theme: Agent Systems.

We're seeking teams building autonomous systems of agents that deliver complete solutions to concrete problems.

Read the full thesis: beta.works/camp
Apply! beta.works/agentsystems
Priority deadline is 12/24 🎄

Scaling agent systems to solve complex tasks can lead to inefficiencies. We introduce asymptotic analysis with #LLM primitives (#AALPs) to model and optimize computational cost as systems grow.
Read our paper: ow.ly/ALHq50W2YaM
Dive into our blog: ow.ly/YuF950W2YaL
#AgentSystems

Discoverable agents completely changes the design of multi-agent systems.

At the @artinet.io, we've been developling architecture that allows #AIAgents to dynamically discover each other based on their capabilities.

Empowering agents to solve evermore complex problems.

#AI #ArtificialIntelligence

From tool calling to multi-agent orchestration—LLMs now interact via the Model Context Protocol (MCP) for scalable, modular workflows. Agents coordinate tasks through JSON-RPC, powering the next generation of AI systems.
#LLM #MCP #AgentSystems #AI #OpenAI #Anthropic

Детальна розбірна архітектура Emergent AI 2026: як агенти приймають рішення та взаємодіють у системі. Технічний гайд для розробників і AI-ентузіастів.
🔗 webscraft.org/blog/yak-pra...
#EmergentAI #AgentSystems #AI

"Decentralized coordination with shared state: A new approach for multi-agent systems explored in kiro-flock open-source reference implementation. 🛠️🤖 #AI #AgentSystems #DecentralizedCoordination"

UC San Diego researchers say multi-agent AI has a memory problem. The fix? Computer architecture: caching, consistency, hierarchy. Old problems, new systems.

#AI #AgentSystems

https://resultsense.com/i/2026-03-19-multi-agent-memory-computer-architecture-perspective

Introduction to AI Agent Development: Multi-Agent Design Patterns and Implementation Guide

#AIagents #multi-agentsystems #LLM
https://singulism.com/en/2026-05-06-ai

https://github.com/agentsystems/agentsystems/

Agent JIT Compilation: Compiling NL tasks to code for web automation. Achieves 10.4x speedup & 28% accuracy gain over sequential LLM agents. #AI #AgentSystems #WebAutomation

https://www.startuphub.ai/ai-news/ai-research/2026/agent-jit-compilation-for-web-automation

Code is no longer just output for LLMs—it's the agent's 'harness.' This framework enables executable, verifiable, and stateful AI agents, from coding assistants to multi-agent coordination. #AI #LLM #AgentSystems

https://www.startuphub.ai/ai-news/ai-research/2026/code-as-the-agent-harness

I just published a small deterministic demo for a bounded weak-signal routing pattern:

Notice → Interpret → Shape → Govern

Weak cues should matter without instantly becoming truth, memory, or action.

#AI #Robotics #CognitiveScience #HCI #AgentSystems

“AI sees itself” is not an architecture.

Mirror recognition, if you want to talk about it seriously, should be modeled as a bounded, governed event under reflection — not as consciousness fanfiction.

New paper:
zenodo.org/records/1901...

Demo Monday.

#AI #AISafety #AgentSystems

Agents are ditching RAG for pure vector search—because memory frameworks already store embeddings. Find out why similarity search is becoming the go‑to retrieval infra for LLM agents. #VectorSearch #AIMemory #AgentSystems

🔗 aidailypost.com/news/agents-...

I am a bit disappointed in the reference information which the AI has generated for our Practical Guide paper. Ann Nowe is a fantastic researcher and contributed a lot to our paper, but I still don't think she deserves to be given my authorship slot in addition to her own! 😊

AI Infrastructure and Agent Systems
Reliability 52% · Impact 64%
https://newshive.geekybee.net/stories/ddf1eec4-a29f-4c00-8760-c8b290d6e296
+1 more updated this hour.
#NewsHive #AIInfrastructure #AgentSystems

A bigger context window does not guarantee better retrieval. Needle-in-a-haystack shows many LLMs miss facts buried in the middle of the context, and agent systems can favour tools listed first over better later options. Better structure beats bigger prompts.
#LLMs #AIEngineering #AgentSystems

The AI ecosystem is targeting effective agents. Insights from the Consortium show that tool use boosts performance, and future breakthroughs might revolutionize AI. #AI #MachineLearning #AgentSystems

Source