AI Weekly Malaysia

Back to items Summaries

A Fireside Chat with Cat and Thariq from the Claude Code team

ID
6422
Status
summarized
Published
21 Jul 2026, 8:54 PM
Fetched
21 Jul 2026, 9:46 PM
Provider
Simon Willison
Category
developer-ai
Original URL
https://simonwillison.net/2026/Jul/21/cat-and-thariq/
Source URL
https://simonwillison.net/atom/everything/

Summary

Score
8.5
Created
21 Jul 2026, 9:46 PM
Tags
Audience
developersvibe_codersai_agent_usersai_ml_learners

What happened

Simon Willison hosted a fireside chat with Cat Wu and Thariq Shihipar from Anthropic's Claude Code team at the AI Engineer World's Fair, covering Claude Code, Claude Tag (a Slack integration), Fable, coding agent security, evals, and tool design. Key takeaways include Claude Tag now landing 65% of product engineering PRs for the Claude Code team, system prompts shrinking 80%, and the finding that adding examples or 'don't do X' lists to system prompts is no longer best practice for newer models like Fable 5 and Opus 4.8.

Why it matters

For developers and vibe coders, this reveals how Anthropic itself uses coding agents at scale — dogfooding internally, relying on automated code review for outer layers, and shifting from micromanaging agent permissions to delegating implementation. The prompt engineering shifts (smaller prompts, fewer negative instructions, fewer examples) directly affect how builders should approach their own agent workflows today.

Discussion angle

How the shift away from example-heavy, negative-instruction-heavy system prompts changes the way Malaysian dev teams should structure their own AI coding agent workflows — and whether the 'delegate more, monitor less' approach is realistic for smaller teams without Anthropic's internal tooling.

Top