Used Claude's new Code Interpreter for a real task for the first time this morning – uploading a 1,800 line CSV file (not one I'm willing to share) and asking for analysis of data over time and theories on why certain trends had happened – Opus 4.1 did an outstanding job with it
@simonw
-
Claude Web App Now Executes Python Code Scripts
By
–
Yes definitely – it only works in the web app right now and you have to turn on the feature in settings (if it's available to your account) Here's an example where I pasted in Python code and told it to "Run this python script and show me the result" https://
claude.ai/share/4d9c603e
-c88d-4b94-bf8b-7b8e24444434
… -
Two and a half years of ChatGPT appreciation
By
–
Right, I've been writing about how much I like the ChatGPT version for 2.5 years now
-
Remote Linux Sandbox System for AI Code Execution
By
–
That's why I'm so puzzled by how they promoted it – it sounds like it's similar MCP file access but it's actually something much more interesting, it's a full remote Linux sandbox system with Node.js and Python and Bash and the ability to install extra packages
-
Naming AI Systems: The Challenge of Finding the Right Identity
By
–
Hah, you've absolutely got me there – so much easier to criticize others than actually come up with a good name myself Claude Analyst perhaps? I get why they're finding this so hard
-
LLM Code Execution Vulnerabilities: Shell Commands and Arbitrary Code
By
–
Haha that answer it gave you isn't actually correct: it CAN run arbitrary shell commands but you have to convince it to use Python's http://
subprocess.run(…, shell=True) You can even get it to run PHP or Deno or Lua if you know what you're doing https://
til.simonwillison.net/llms/code-inte
rpreter-expansions
… -
Browser-based JavaScript feature execution implementation notes
By
–
I do have another version of this feature which executes JavaScript and runs in the browser – here are my notes on that one
-
Claude API Code Execution Tool Features Compared
By
–
That one isn't a feature of the http://
Claude.ai consumer-facing app, it's only available to users of their API https://
docs.anthropic.com/en/docs/agents
-and-tools/tool-use/code-execution-tool
… I expect the new user-facing feature uses the same underlying architecture though -
Transparency in LLM Usage: Sharing Prompts and Transcripts
By
–
I generally don't use custom rules because I don't like to get an experience that's different from the "default" everyone else gets – it makes it harder to write about things. When I write about Codex or GPT-5 or other LLMs I always share my exact prompts and transcripts
-
Server-side code execution upgrade: Python and Node.js
By
–
Their previous one ran JavaScript code in your browser – this is an upgrade to that which runs Python and Node.js code on the server instead. Here are my notes on the old one: