Codex is moving from a coding tool to an AI agent for work. What that means, and how to brief it well.
Why AI agents need memory, tool management, identity, logging, security rules, and a control layer around them.
Anthropic introduces five effort levels for Claude. Developers decide exactly how long the model thinks, and what it costs in time and tokens.
Replit founder Amjad Masad laid off half his team. A small group in a back room built a coding agent. Twelve months later: a hundredfold jump.
Anthropic's new Memory and Dreaming for Claude Managed Agents turn AI agents from one-shot helpers into systems that learn between sessions.
Fiona Fung leads the Claude Code team. When AI took over the coding, every old team norm had to be rewritten. Here's what survived.
YC's president rebuilt Posterous in five days for $200 with Claude Code. His formula: thin harness, fat skills, and never skimp on tokens.
DeepMind CEO Demis Hassabis holds firm on AGI by 2030: drug discovery in days, new sciences from AI simulations, and a new view of reality.
Karpathy separates vibe coding from agentic engineering, and explains why you can outsource your thinking but not your understanding.
Boris Cherny, the creator of Claude Code, explains why he hasn't written a line of code in 2026, and what AI agents still haven't solved.
Claude Code tells the story of how Glenn builds a Norwegian accounting system from scratch, and what sets an AI agent apart from a chatbot.
a16z, Box, and Steven Sinofsky on why most enterprise AI projects fail, why agents hit a wall at integration, and why jobs are still safe.