On the plus side, at least an exact string match on the system prompt means I don't have to worry that if my legitimate Claude Code usage occasionally mentions the string "OpenClaw" I might get hard-to-debug intermittent failures
@simonw
-
Anthropic’s Claude Max plan filtering on system prompts concerns
By
–
I wasn't particularly offended when Anthropic said that their Claude Max plan was reserved for their own harness – I bought the "we've optimized ours for cost" argument – but filtering on system prompts feels a step too far to me
-
Anthropic’s discriminatory billing based on system prompt content
By
–
Billing different based on text contained in the system prompt is a really bad look Peter Steinberger 🦞 (@steipete) Anthropic now blocks first-party harness use too 👀 claude -p –append-system-prompt 'A personal assistant running inside OpenClaw.' 'is clawd here?' → 400 Third-party apps now draw from your extra usage, not your plan limits. So yeah: bring your own coin 🪙🦞 — https://nitter.net/steipete/status/2040811558427648357#m
-
uvx Workflow Integration and Tool Installation Concerns
By
–
Yeah I do worry about that, especially given how key to my workflow uvx has become I'm not sure telling people to "pip install" or "uv tool install scan-for-secrets" and then "scan-for-secrets –help" is materially different though
-
AI Healthcare Messages from Hospital Desert Areas
By
–
Are those 600K weekly messages from hospital deserts classified as healthcare related?
-
README-Driven Development with Claude Code for Tool Building
By
–
I built this one using README-driven-development: I hand crafted a detailed README describing exactly how the tool should work… then dumped that into Claude Code and told it to build it gisthost.github.io/?d4b1a398…
-
Intelligence Services Can Easily Penetrate Technology Secrecy Systems
By
–
I expect the secrecy around them is easily penetrated by any remotely competent intelligence service – thousands of people must be in the loop, and they will have signed NDAs, not military-grade secrets laws
-
New Blog Tag Tracking AI-Powered Security Research Stories
By
–
Started a new tag on my blog to track stories about AI-powered security research, which is very much having a moment right now – 11 posts so far already simonwillison.net/tags/ai-se…
-
Coding Agents Require Significant Cognitive Load and Burnout Risk
By
–
Just noticed this has had 1.1m views now, which explains why I starting to see some less informed reactions to it starting to crop up now it's broken out of purely tech Twitter https://t.co/XFn8uYb6os
— Simon Willison (@simonw) 3 avril 2026Just noticed this has had 1.1m views now, which explains why I starting to see some less informed reactions to it starting to crop up now it's broken out of purely tech Twitter Lenny Rachitsky (@lennysan) "Using coding agents well is taking every inch of my 25 years of experience as a software engineer, and it is mentally exhausting. I can fire up four agents in parallel and have them work on four different problems, and by 11am I am wiped out for the day. There is a limit on human cognition. Even if you're not reviewing everything they're doing, how much you can hold in your head at one time. There's a sort of personal skill that we have to learn, which is finding our new limits. What is a responsible way for us to not burn out, and for us to use the time that we have?" @simonw — https://nitter.net/lennysan/status/2039845666680176703#m
-
Claude Code Feature Limitations: Interactive Prompting Gap
By
–
The feature set is the problem – the model and harness are great, but in Claude Code for web I can do things like prompt the model while it's working