Version 1

Simon Willison published two posts on May 4, 2026, using Claude Code as an agentic rapid prototyping tool to explore and demonstrate emerging technical developments. In the first, he built a browser-based interactive playground that runs a WASM-compiled subset of Redis, allowing users to experiment without a local install with a significant in-progress Redis feature — a native array data type proposed by Salvatore Sanfilippo (antirez) in a PR that adds over 18 new commands[1]. The playground makes the new API immediately accessible in the browser, exemplifying Willison's use of Claude Code to lower the barrier to exploring nascent technology.

The second post zooms in on one of those new Redis commands: ARGREP, which enables server-side regex grep over array values using the TRE regular expression library recently vendored into Redis by antirez[1][2]. Curious about TRE's security properties, Willison used Claude Code to build an experimental Python binding to TRE via ctypes and tested it against pathological ReDoS inputs[2]. His finding: TRE handles ReDoS attacks significantly better than Python's standard re module, primarily because it does not support backtracking — the mechanism most commonly exploited in regex denial-of-service attacks[2].

Across both posts, a consistent framing emerges: Claude Code is positioned not as a code generator for production systems but as an agentic collaborator for rapid prototyping — quickly building just enough to probe interesting technical territory and make it tangible for an audience. The Redis playground and TRE binding are explicitly experimental artifacts whose value lies in demonstration and exploration rather than deployment. The background Wikipedia articles included in this thread (on Llama, 3D printing, Memex, regex engine comparisons, and TRE computing) carry no direct claims of their own and appear to be reactive context documents surfaced by the pipeline as supplementary background.