Comment by ManuelKiessling

7 days ago

Thanks for looking into it.

While I would have hoped for a better result, I'm not surprised. In this particular case, I really didn't care about the code at all; I cared about the end result at runtime, that is, can I create a working, stable solution that solves my problem, in a tech stack I'm not familiar with?

(While still taking care of well-structured requirements and guard rails — not to guarantee a specific level of code quality per se, but to ensure that the AI works towards my goals without the need to intervene as much as possible).

I will spin up another session where I ask it to improve the implementation, and report back.

2 comments

ManuelKiessling

necovek 6 days ago

I'd definitely be curious to see if another session provides higher quality code — good luck, and thanks for taking this amicably!

ManuelKiessling 4 days ago

I did another session with the sole focus being on code quality improvements.

The commit with all changes that Cursor/claude-3.7-sonnet(thinking) has done is at https://github.com/dx-tooling/platform-problem-monitoring-co....

As you can see, I've fed your feeback verbatim:

  I have received the following feedback regarding this codebase:

  "The premise might possibly be true, but as an actually seasoned Python developer, I've taken a look at one file: @utils.py. All of it smells of a (lousy) junior software engineer: from configuring root logger at the top, module level (which relies on module import caching not to be reapplied), over not using a stdlib config file parser and building one themselves, to a raciness in load_json where it's checked for file existence with an if and then carrying on as if the file is certainly there..."

  I therefore ask you to thoroughly improve the code quality of the implementation in @src   while staying in line with the requirements from @REQUIREMENTS.md, and while ensuring that the Quality Tools (see @makefile) won't fail. Also, make sure that the tests in folder @tests  don't break.

  See file @pyproject.toml for the general project setup. There is already a virtualenv at @venv.

You can watch a screen recording of the resulting Agent session at https://www.youtube.com/watch?v=zUSm1_NFKpA — I think it's an interesting watch because it nicely shows how the tool-based guard rails help the AI to keep on track and reach a "green" state eventually.