Boris Cherny

Claude Code adds /checkup, a housekeeping command that removes unused skills and MCPs, dedups and restructures CLAUDE.md files, turns off slow hooks and updates the CLI.

LLM-produced bugs have shifted from off-by-ones to system design, UI usability and missing context. Adversarial code review catches many of them, and can be as little as a one-line prompt.

Anthropic confirms a data-retention change for its Mythos-class models: customers own and control their data and Anthropic retains none, shipping this fall. Aimed at enterprises whose privacy and compliance rules currently block frontier-model deployment.

Anthropic's Boris Cherny says Claude's chat interface and its Cowork agentic-work product are merging into one experience that carries context across sessions, rolling out gradually.

Boris Cherny previews Function Hooks, an unreleased extensibility layer for Claude Code, alongside demo videos from the ClaudeDevs account. The design is still open: Anthropic is gathering feedback on a GitHub issue before committing to the interface.

Shopify's CTO discusses unlimited token access for all engineers, defaulting to the largest model, a digital twin of the company for predicting merchant growth, and a barbell token strategy for getting the most from Claude.

Claude Code is Anthropic's AI coding agent that lets developers ship code through natural language instead of writing it by hand. It supports agentic workflows where you can dispatch multiple PRs at once, even from a phone.

Boris Cherny highlights the result he finds most exciting in the Opus 5 system card: buried on page 73, it is Anthropic's least prompt-injectable model yet across injection evals and red teaming.

Claude Code creator Boris Cherny at Startup School on how the tool came to be, what cutting 80% of its prompt did, and what building products means when the underlying model capabilities keep accelerating.

Claude Code is Anthropic's AI coding agent that lets developers build software through natural language rather than a traditional IDE. It can autonomously write, edit, and reason about code — even building entire applications like Cowork in 8 days.

Anthropic's Boris Cherny says OpenAI's newest model now matches Gemini Flash and Claude Opus 4.8 on prompt-injection resistance, claims Claude solved it first two months ago, and argues labs still need more alignment work as agents take on higher-stakes tasks.

Claude Tag in Slack has moved to Fable 5.1, shown building a leadership deck from a metrics spreadsheet and scattered Slack context while flagging a vendor report that contradicts the numbers. Availability is limited to Slack Team and Enterprise plans.

Boris Cherny on Claude Code going from a terminal prototype to 4% of public GitHub commits in a year, the counterintuitive product principles behind it, and what happens to the profession once coding is solved.

Claude Code is an agentic coding tool that runs in your terminal, letting developers delegate coding tasks, verification, code review, and CI automation to an AI agent. It supports advanced workflows like auto mode, agent orchestration, voice control, and remote operation across entire codebases.