This matches what's happening in agent systems beyond research too. You would find similar patterns in Claude Code as well. I think the best agent architectures keep converging on thin orchestrators that delegate instead of doing the work themselves.
AGENTS
-
Qwen 3.5 27B NVFP4 runs 5 agents full context under 20GB VRAM
By
–
Qwen 3.5 27B in NVFP4 w/ full context taking less than 20GB VRAM You can basically run like 5 agents w/ full context on a single RTX PRO 6000 like this, and they'd be so fast Tell me I didn't tell you this was gonna happen
-
GIPO: Gaussian Importance Sampling for AI Learning Without Forgetting
By
–
What if your AI agent could learn faster from older data without forgetting or crashing? Researchers from Sany Group present GIPO (Gaussian Importance Sampling Policy Optimization). Instead of hard clipping, it uses a soft Gaussian trust weight to gently limit extreme updates
-
Glean Prompt Library: AI Agents for Enterprise Search
By
–
Glean Prompt Library! @glean Glean’s platform seamlessly integrates with a wide array of enterprise systems such as email, intranet, cloud storage, and database by providing a powerful, unified search interface with AI Agents that effortlessly retrieves information from all
-
DeepMind LLMs Lean Proof-Search Agents Solve Open Problems
By
–
This new DeepMind research turns LLMs into Lean proof-search agents, so every step must compile and the final proof is mechanically verified. Under this setup, they solved 9 open Erdős problems, proved 44 OEIS conjectures, and helped advance actual research in optimization,
-
Agent Behavior: Each Line Must Change Agent Actions
By
–
The fix is making every line answer "what does the agent do differently because of this?" If a line doesn't change behavior, it's decoration.
-
Self-Play SWE-RL: Training Superintelligent Software Agents
By
–
The bottleneck for coding agents may not be code. It may be experience. A new ICML 2026 paper introduces Self-play SWE-RL (SSR): Toward Training Superintelligent Software Agents through Self-Play SWE-RL The question is simple and profound: How do you train software agents
-
Managing a Legion of AI Agents Daily
By
–
Sorry babe I can't read books any more, I'm spending all of my waking hours managing a legion of agents.
-
Plugins affect empty state of new menu via datasette-agent
By
–
Plugins can also affect the empty state of the new menu – the latest datasette-agent adds a form for kicking off a new agent conversation – live demo (if you sign in with GitHub) on https://t.co/hPpyS2cumO pic.twitter.com/tmWTmX7ZMc
— Simon Willison (@simonw) 25 mai 2026Plugins can also affect the empty state of the new menu – the latest datasette-agent adds a form for kicking off a new agent conversation – live demo (if you sign in with GitHub) on https://
agent.datasette.io -
Chrome Extension Syncs Multi-Platform AI Conversations
By
–
How to Use Memory～ Someone turned the Chrome browser into an inbox for all your AI conversations. Conversations from ChatGPT, Gemini, Claude, DeepSeek, and Grok across all five platforms are synced in real-time to local markdown files. There's an agent hanging in the sidebar