Continuation of `Having "blessed" 3rd party libraries may make D more popular` DIP thread

[Richard (Rikki) Andrew Cattermole]

On 08/07/2025 10:03 AM, WraithGlade wrote:
> Tangentially from my just prior post, I have a question:
> 
> What is the D language team's stance on AI derived from data harvested 
> from the internet (or anywhere else) without regard to that data's IP or 
> moral rights?
> 
> I personally find the advent of plagiarism/theft based AI systems to be 
> repugnant and unpalatable to such a great degree that if such AI were to 
> pervade all tech companies as mandatory use (such as how Microsoft 
> allegedly now requires its use by all employees, very tellingly)  then I 
> would literally rather leave the entire tech industry than ever use it, 
> even despite devoting my lifetime to programming and mathematics/logic 
> so far and even if leaving the industry doomed me to a life of poverty. 
> I'm not even kidding about that.
> 
> I am still unsettled as to what language and ecosystem to build on, so I 
> am trying to find any pieces of information that could conceivably 
> clearly decide what choices are or aren't a fit for me.
> 
> Anyway, I need to go eat now. See you guys later and have a good day/night.

Given that we do not have an official IDE, let alone an LLM assistant, D 
as a project can IMO ignore the implications and the politics for LLM 
creation. This is much bigger than us, and no decision we could make 
would be accepted more broadly. This covers politics in general, as long 
as its not giving us problems our concern is more with staying out of it.

What we do have is contributions from people, and they may use LLM's.
This we have a stance on. I brought this to the monthly meeting a couple 
of months ago. Our decision was that LLM's are a tool, but what we want 
in contributions is from Humans. Humans must make the decisions, tools 
are just what we use to get there.

The reason this came up is due to a couple of PR's that were full of LLM 
generated content, and the comments within the PR itself included LLM 
content that was not understood. The people involved did not understand 
what they were attempting to contribute.

As a result of this, I did this PR: https://github.com/dlang/dmd/pull/21364