claude ai-agents developer-tools code-generation

How to stop Claude from saying load-bearing

jola.dev

July 14, 2026

2 min read

🔥🔥🔥🔥🔥

69/100

Summary

The MessageDisplay hook allows users to modify Claude's vocabulary, including replacing phrases like "load-bearing seams." A Python script can be created to implement these vocabulary changes through specific replacements.

Key Takeaways

A script can be created to replace specific phrases in Claude's output, such as changing "load-bearing" to "cooked."
The script utilizes Python and regular expressions to perform text replacements based on a predefined dictionary of phrases.
Users can implement the script by placing it in the appropriate directory and configuring their Claude settings to use the new hook.
Hooks in Claude load at startup, requiring users to start a new session to activate the changes.

Read original article

View story →

Coverage & discussions

Hacker News · jola.dev

How to stop Claude from saying load-bearing ↗

Community discussions

Hacker News discussion ↗

Historical count: 5 · 5 saved in sample · Historical sample

Analysis ready

Community Sentiment

Negative

Positives

Some users appreciate Claude's unique phrasing, noting that everyone has their own voice, which is usually acceptable until it's echoed by millions of generated tokens.
One commenter enjoys experimenting with Claude's output to make it sound more human, suggesting that there's still potential for creativity in AI-generated text.

Concerns

Claude's repetitive speech patterns are driving people crazy, making it sound like a frustrating project manager rather than a helpful assistant.
Commenters are concerned that LLMs are getting stuck in predictable verbal tics, which diminish their usefulness and make interactions tedious.
The focus on RLHF seems to have stripped models of their ability to convey unique voices, leaving users feeling like they're conversing with a soulless machine.