docs: added a declaration of GenAI usage to the AI-POLICY.md file

Author: tobixen

AI-POLICY.md

@@ -2,7 +2,12 @@
 
 ## Read this first
 
-The most important rule: Be honest and inform about it!
+The most important rule: be honest and inform about it!
+
+Also: keep a log of the prompts used - prompts may be included in the
+git commits.
+
+## Transparency matters
 
 If you've spent hours, perhaps a full day of your time writing up a
 pull request, then I sort of owe you something.  I should spend some
@@ -26,10 +31,10 @@ explain in details why I'm rejecting the pull request.
 ## Bugfixes are (most often) welcome
 
 It's fine to ask the AI for help to analyze a bug and create a fix for
-it.  By discovering the bug, reproducing it and testing it you're adding
-real value to the project - but be transparent about AI usage and I
-expect that you will not break down and cry if I decide to reject the code
-changes.
+it.  By discovering the bug, reproducing it and testing it you're
+adding real value to the project - just be transparent about AI usage
+and do not take offence if the code changes are rejected, or completely
+rewritten.
 
 ## General rules
 
@@ -69,3 +74,19 @@ changes.
   as of 2026-02, and I can hardly see cases where the AI would break
   the Code of Conduct, but at the end of the day, it's **YOUR**
   responsibility that the contribution follows those guidelines.
+
+## Declaration of "GenAI" usage
+
+After receiving some really shitty pull request generated by Claude Code, the maintainer started playing with Claude Code himself in the end of 2025.  I wrote a blog post summarizing my experiences on https://www.redpill-linpro.com/techblog/2026/03/20/from-luddite-to-vibe-coder.html
+
+Releases 2.2.6 - 3.2.0 has been heavily assisted by Claude - which is pretty obvious when looking into the commit messages.  My experiences has been mixed - sometimes it seems to be doing a better and faster job than me - but despite (or because of?) using Claude extensively, the 3.0-release got quite much delayed, I spent much more time on it than estimated.  There were quite some bug reports pouring in after the 3.0-release - but I don't think it's related to the AI-usage.  The CHANGELOG for v3.0.0 says it quite clearly: "there are massive code changes in version 3.0, so if you're using the Python CalDAV client library in some sharp production environment, I would recommend to wait for two months before upgrading".
+
+Post-3.2.0 and until further notice I will try to go more back to the old ways for doing the "core development tasks" - mostly manually.  This allows me both to have full control of the process as well as to maintain my brain cells and coding skills.  GenAI-tools should be used mainly for improving quality, not speeding up the development.  This includes lots of stuff:
+
+* Dealing with test framework.  It's hard work, even when using Claude.  Thanks to Claude I've now been able to put up an extensive "battery" of test servers that I'm checking regularly towards.  This is very important, most of the bug reports I get in is due to exotic servers and server setups doing weird things - the more servers I have for testing every release, the less troubles will be discovered downstream.
+* Other CI-related frameworks and boiler plate ... like, now there is even QA in place for commit messages etc before I push my git commits out from my laptop.  It increases quality, although being quite outside the "core business" of the CalDAV library.  Doing it manually would have stolen lots of valuable time that could have been used for coding.
+* Writing up test code.  I've always thought that "test driven development" is a good idea (write test code first, then the logic), but it's quite often both tedious and difficult.  Claude can make them really fast.  There is a risk that the tests are optimized for passing rather than for finding bugs - this is nicely circumvented by having Claude creating the tests before the code is made, as well as asking Claude to make test code exposing the bugs rather than doing ad-hoc debugging.
+* Code reviews.  The more "eyes" looking into the software, the better - it seems Claude is equally good at spotting the problems and mistakes in my code as I'm on spotting the problems and mistakes in the code Claude generates.
+* Debugging.  It's easy to get stuck and spend tons of time on debugging - sometimes (but not always) Claude can find them easily.
+* Various mundane and tedious work (i.e. "I left a TODO-note in the code over there, could you have look into it and eliminate it?").
+* Development of the caldav-server-checker tool - writing up checks to discover various server issues may be really tedious and time-consuming.  Claude needs some supervising also while doing this work, but the alternative would probably be to have half as many checks.  Knowing and verifying the server behaviour is extremely important.