almost entirely –yolo, usually in a remote sandbox, often locally and not in a sandbox, still trying to find a local sandbox that I like
@simonw
-
Sandboxing practices for safe AI code development
By
–
Git! But yeah, that's why I mostly YOLO in Claude Code for web, then it's in a container where it can't damage anything I don't yet have a local sandboxing habit that I stick with, but I really need one
-
YOLO Users: Testing AI Systems in Sandbox Environments
By
–
Follow-up for –yolo users: are you YOLOing in a sandbox?
-
AI coding agents: security flags and permission handling
By
–
Coding agent users: do you run with –yolo (Codex) or –dangerously-skip-permissions (Claude Code) or equivalent?
-
Qwen Model Highlights AI Capabilities Analysis
By
–
One of the Qwen models did call that out a while back https://
simonwillison.net/2025/Aug/10/qw
en3-4b/
… -
AI-Assisted Coding: Typing Less, Programming More
By
–
I don't see this as me not coding anymore – I see it as me not typing the code out myself
-
Claude Code Opus 4.6 Builds Library One-Shot Prompt
By
–
With Claude Code and Opus 4.6 I practically one-shot prompted this library which I've been wanting to build (and procrastinating on) for a couple of years
-
Positive Results with Private Animal AI Prompts
By
–
It's been fantastic agains the private animal prompts I've tried
-
GPT-5.2 Model Context Window Specifications and Documentation
By
–
I tend to go by the model listings like this one which don't say anything about thinking v.s. non-thinking getting different context windows https://
platform.openai.com/docs/models/gp
t-5.2
… -
ChatGPT Plan Comparison Grid Confusion Explained
By
–
Anyone else confused by the new ChatGPT plan comparison grid? Here's an annotated screenshot: