Very Good Read #0
This is a pilot for a new format of posts in which I highlight some good reading material I found while perusing the interwebs, as we all know it is getting harder and harder to find worthwhile reading articles with the flood of LLM generated slop that we are having to deal with nowadays, but not all is lost, there are still very smart and knowledgable people out there putting out some very good and informative reads.
This time I would like to invite y'all to read "To Gen or Not To Gen: The Ethical Use of Generative AI"[1] by Johannes Link and Jakob Schnell, the former also being the author of Jqwik[2] a property-based test engine for Java that gained some attention lately for commits that basically added a harmless prompt injection[3] so coding agents would simply stop running after reading the test outputs which, quite frankly, was a very smart thing to add, I was both laughing at the people complaining, some even saying the codebase now had malware in it and needed to be reported[4], and in awe at the creativitiy and geniosity of the mind behind that change at the same time when I found out about this, but personal sentiments aside, this read is a very good one IMO because it is an excellent summarization of how and why the usage of generative AI is extremely unethical and harmful, it goes as far as enumerating the steps that would have to be followed if we ever wanted to make the current LLM craze to become ethical. While most people in software development refer to this subject from a quality or copyright POV, this article is the first one I've found so far that looks into all this somewhat extensively with a heavy focus on ethics, which aligns completely to how I see this situation as well, because I do firmly believe we need to talk about it in this way I think it's very worth checking out.
PS: If you want to know more about the small scale scandal that happened around Jqwik, Johannes made a blog post[5] explaining everything about it.