The point of the summary is also to explain "why" something was done, most Claude-generated PR descriptions I've been seeing go through the "what" and "how" but if the human-in-the-loop didn't care to precisely describe the "why" it is just an English version of the changes made in the code... I can just read the code for that, give me the reasons behind the diff and I'm a happy camper.
If you have a large PR the existence of a good summary on "what" changed can help you to make a better review.
But I agree with you, when reading PR descriptions and code comments I want a "why" not a "what". And that is why I think most LLM-generated documentation is bad.
Current social consensus is that copyright exists and one can only use software on conditions stated in license. Thus, proprietary and copyleft have same protection.
Another possible consensus would be that copyright don't exist, and anyone can copy proprietary or copyleft work and improve it. Nobody would be harmed in such situation, original author still have its copy. I would have no problem with such state - but it must be same for everyone, not just FOSS.
If I release something as MIT or Apache, all I want is some credit, either for my own self-satisfaction or for resume fuel.
If a library I wrote was used by BigCo, then I could point to their license file and mention that in a job interview or something. If they have Claude generate something based on my code, they don't put it into their license, I don't get the resume fuel, and my work is unrewarded.
I have gone back and forth about how I feel about AI training on code, and whether I think it's "theft", but my point is that the original code being available is kind of missing the point.
The bitrate of the PCM is determined based on how quickly you can write a byte to the register. The fastest you could write general data is once every 6 cycles, which gives ~298 MHz of sample rate, so 44.2 kHz is easily doable if that's all you want to do with the CPU.
