@mehaisi

When AI Agents Get Corporate Cards: The Spend Authorization Problem No One Solved

7h ago · 4 min read · Ramp just gave AI agents their own corporate cards. The real story isn't the card—it's the authorization layer underneath. Here's what's happening: Ramp's new Agent Cards let AI agents spend money without ever seeing a card number. The agent calls an...

HTTP 402 Is Finally Becoming What It Was Always Meant to Be

11h ago · 4 min read · Stripe and Tempo just repurposed a 30-year-old HTTP status code into a payment protocol for AI agents. The Machine Payments Protocol turns 402 "Payment Required" into a full authentication scheme for autonomous transactions. Here's what that means—an...

The Fraud Infrastructure Gap for Autonomous Agents

11h ago · 3 min read · When Claude can autonomously generate a funded prepaid Visa card and complete a purchase without human intervention, the question stops being "can AI agents spend money?" and starts being "who built the infrastructure to catch them when they go wrong...

The Chatbot Is Dead: Why Anthropic Is Building an AI OS, Not a Better Assistant

19h ago · 5 min read · The Chatbot Is Dead: Why Anthropic Is Building an AI OS, Not a Better Assistant Anthropic just shipped something that didn't make many headlines. Claude's mobile update now embeds live, interactive instances of Figma, Canva, Amplitude, and other ente...

The .claude/ Directory: Why Your AI Agent Needs a Project README

19h ago · 5 min read · The .claude/ Directory: Why Your AI Agent Needs a Project README Boris Cherny built Claude Code at Anthropic. He runs five parallel Claude instances from his terminal, each one capable of one-shotting a full feature implementation. His secret isn't a...