Posts tagged 'Anthropic'.
Claude Code's April 2026 update adds an interactive Google Vertex AI setup wizard, a real-time background-script monitoring tool (Monitor), Linux PID namespace isolation for subprocess security, and a 60% speed boost on the Write tool. Everything developers and educators need to know.
In March and April 2026, Anthropic released Auto Mode (autonomous execution) and Claude Managed Agents (a managed agent API) for Claude Code. This analysis covers what it means for AI coding to shift from assistant to autonomous agent β and what that means for developers and educators.
The April 2026 Claude Code update introduces /powerup interactive tutorials, MCP connection improvements, PreToolUse defer hooks, and PermissionDenied events. Here is how your AI coding workflow changes.
Claude Code v2.1.91 expanded the MCP tool result limit to 500,000 characters. v2.1.92 introduced the /cost command for real-time per-model, per-cache cost analysis. Two updates released in the first week of April 2026 that meaningfully change AI development workflows.
In the first week of April 2026, two seismic events rocked the AI world. Anthropic accidentally exposed a secret model called Claude Mythos β described as posing "unprecedented cybersecurity risks" β while Harvard formally adopted vibe coding as part of its curriculum. An edtech CEO breaks down what both events mean for education.