#mcp

Frontier models now beat the human baseline on OSWorld-Verified — but the benchmark just got rebuilt, and the architecture quietly shifted off pixels.

MCP went stateless, a wave of coding-agent RCEs landed, and a new benchmark measured reward hacking — the three properties that make an agent useful all became liabilities.

The week of May 19 separated the agent loop from tool execution. Whoever hosts the loop now owns your latency, reliability, and lock-in.

Skill cards, self-hosted sandboxes, MCP tunnels, computer-use verifiers, and a Five Eyes warning all landed in twenty-one days. The boring perimeter around capable agents finally has shape.

Twenty minutes of tabs vs. five minutes of prompt. The traditional web wasn't designed for humans — it was designed for mice. The agent-native web is quietly dismantling the parts that never made sense.

Anthropic just shipped 10 financial services agent templates. The interesting part isn't the templates — it's the three-layer architecture quietly becoming the standard for enterprise agents.

When one agent isn't enough, how do they talk? A2A for peer-to-peer delegation, MCP for tools, and the durable message-bus patterns underneath multi-agent systems in 2026.

Tools are what turn a chatbot into an agent. The 2026 stack for tool design — small scoped tool sets, structured arguments, MCP servers, and the failure modes you only see in production.

AI agents are drowning in tools. The more APIs you connect via MCP, the worse your agent performs. Here's why, and what Code Mode changes.

Past ~30 tools, agent reliability falls off a cliff. Past ~100, it's chaos. Here's the actual engineering — RAG-over-tools, semantic routing, dynamic loading, and namespacing — that production teams ship to stay sane.

Fourteen months in, the Model Context Protocol stopped being a curiosity and started being plumbing. Here's what changed over the holidays — registries, governance, and the first scaling pains.