Anthropic AI Updates: April 18, 2026

1. Anthropic Labs Launches Claude Design, a Prompt-to-Prototype Tool

Anthropic. Claude Design is a new research preview from the Anthropic Labs group that turns prompts, images, documents, or codebase context into polished designs, prototypes, slides, and one-pagers. Powered by Opus 4.7, it applies a team’s design system by reading repo and design files so output stays on brand, supports inline comments and custom adjustment controls for refinement, and exports to Canva, PDF, PowerPoint, and standalone HTML. It ships to Pro, Max, Team, and Enterprise subscribers, and Figma shares dropped more than 7% on the news. Source

2. Amodei Meets Trump Chief of Staff to Reset Pentagon Standoff Over Mythos

Anthropic. Dario Amodei sat down with White House Chief of Staff Susie Wiles to try to defuse the administration’s fight with Anthropic after the company refused to let Claude power drone targeting or mass domestic surveillance. The meeting was framed as productive, with both sides discussing collaboration on the new Mythos model; a federal judge has meanwhile ruled that the administration’s attempt to ban and label Anthropic is likely illegal. Parts of the intelligence community and CISA are already testing Mythos under Project Glasswing. Source