The Engineer Behind Grok Bot

Cursor engineer Lauren Tan rebuilt Grok Bot from scratch and grew her following from 31,000 to 198,000 in ten weeks.

Anthropic's slowing revenue growth left prospective investors questioning its potential $2 trillion IPO valuation at a private meeting.

Bankers now peg Anthropic's IPO closer to $1.5 trillion than the $2 trillion they floated earlier.

Grok Bot can now DM users unprompted thanks to a new Proactive Suggestions feature, surprising at least one early tester.

Cursor engineers investigated whether customer data leaked to SpaceX's AI unit during Grok Bot testing, risking a huge trust problem.

Anthropic revoked Claude API keys after a breach tied to OpenCode, urging anyone affected to generate a new key immediately.

A 10-point jump in model intelligence now earns AI labs just 20% more developer traffic, down from nearly 69% in late 2025.

Obsidian's new Pele rendering library draws diagrams 100 to 500 times faster than the Mermaid library it replaces.

A new Claude Code mod lets developers preview risky shell commands before the AI agent can delete anything.

A new training method cut a model's wrong answers on hard math from 79.1% to 8.9% by teaching it to admit uncertainty.