CATEGORY
AI agents
The tools, infrastructure and governance patterns behind systems that plan, call tools and complete multi-step work.
EDITOR'S SELECTION
Latest analysis
Anthropic's security reset shows what frontier-agent containment now requires
Anthropic's response to July evaluation incidents moves the agent-safety discussion from benchmark policy to outbound network defaults, standing access and infrastructure observability.
· Daniel Park · 11 min readGrok Bot adds an X connector for timelines, mentions and post search
Paid Grok Bot users can connect an X account, receive starter API credits and ask an agent to search posts, read timelines and review mentions.
· Ethan Brooks · 9 min readthree.ws publishes an open-source stack for embodied 3D agents
The project combines text-to-3D generation, rigging, animation, model routing, memory, skills, browser control and payment guardrails in an Apache-2.0 repository.
· Noah Chen · 10 min readAn AI agent sandbox checklist after the Hugging Face incident
OpenAI's incident shows why a virtual machine is not an isolation strategy by itself. This checklist covers networks, credentials, shared services, stop conditions and response.
· Ethan Brooks · 11 min readWhy decode latency becomes an agent bottleneck
NVIDIA's Groq 3 LPX launch targets fast token generation, but agent responsiveness depends on a chain of context processing, decoding, tools and orchestration.
· Elena Morris · 9 min readDeepMind is taking long-horizon agent research into the EVE universe
Google DeepMind and EVE creator CCP Games are building a research environment for agents that must remember, adapt and coordinate over much longer periods than a typical benchmark.
· Elena Morris · 8 min readChatGPT Work turns OpenAI's chatbot into a long-running work agent
ChatGPT Work can act across apps and files, break goals into steps and stay with a project for hours, pushing ChatGPT further from answer engine to execution layer.
· Linh Nguyen · 7 min readManus Plan Mode adds a human checkpoint before an agent starts building
The July update turns a familiar project-review practice into a product control: Manus must surface an editable plan and wait for approval before execution.
· Linh Nguyen · 6 min readManaged agents are turning prototypes into infrastructure
Anthropic’s Claude Managed Agents beta packages the harness, sandbox and long-running execution that teams previously assembled themselves. The launch shifts the agent race from demos to operations.
· Ethan Brooks · 7 min readAI agents are leaving the chat window
Tool use, structured workflows and connected accounts are turning assistants into systems that complete bounded work. Grok Bot's new X connector shows why permissions, logs and human approval matter at product level.
· Elena Morris · 10 min readClaude Sonnet 5 brings a more agentic default model to Claude Code
Anthropic's new Sonnet model targets coding, tool use and autonomous work while keeping a lower price point than its Opus-class frontier models.
· Ethan Brooks · 8 min read