Nope, to me it looks like you're coding agent has a strong 🍚 background.

🤭

Sir, typo in the first link 😬
should be: www.yaymemories.com/en

nostalgy? 🥲

Nice 😂 what I do, is asking claude to add a new rule to the claude md file.
Example:
```
No comments in code: Use descriptive function and variable names instead of comments. Code should be self-documenting through clear naming conventions
```
It help a lot.

imho a nice feature of the AI agents is to ask them, after you write your business logic, to analyse the code and ask for feedback.
The agent can catch issues, and suggests overblown test suite. 3/4 of the tests make sense, leave them and feel fairly safe.

> use claude to create my posts (?)

You tell me you ask ai to code and then write a blog post about it? :D

All persons are more then welcome.
"There is no test like production." 🚀 😜

Yes. It helps a lot with not too complex stuff (in terms of code generation), also helps with understanding a flow of the code. I found it useful a few times, when had to analyse what is happening in classes with many lines and responsibilities.

> let it do its thing alone

no way! I still take some time to see what has been generated. If it follows project's patterns, if the code make sense. At the end it has to support us, not replace because it's not possible at the moment. Unless you enjoy closed-eyes programming 😂

Yes, worth giving it a try. Once sceptical, now it's helping me with all boring and repetitive code generations: controllers, repos, simple services, patterns glue code, tests.

One should use it to move faster and progress easier with the domain logic.

Not to show off with numbers or percentages.