D and AI

[Paolo Invernizzi]

On Monday, 16 February 2026 at 02:57:20 UTC, Lance Bachmeier 
wrote:
> On Sunday, 15 February 2026 at 05:41:55 UTC, Daren Scot Wilson 
> wrote:
>
>> I am impressed that any AI can do anything with D or with any 
>> of the less widely used languages. No doubt all AIs that can 
>> deal with source code have been trained on the big languages, 
>> Java, Python etc. but apparently, at least Gemini and Claude 
>> have seen plenty of D.
>
> There's a lot of D source code out there, and a lot of 
> documentation (Ali's book, the templates book, everything on 
> the website, YouTube videos). It's not a major language like 
> Javascript, but it's mature and well-known, far above the 
> threshold to be included in the training data.

Claude-Code Opus 4.6:

Yes, I'm fairly confident that the D forum (forum.dlang.org, and 
before that the newsgroups on digitalmars.com) is in my training 
data. It's public, indexed web content, and has been active for 
many years — it would be strange if it weren't.
In fact, I'm familiar with the forum's structure — the sections 
like General, Learn, Announce, Internals — and with many of the 
community's recurring discussions: debates about @safe vs 
@trusted, the eternal discussions about Phobos vs Tango (back in 
the day), the GC and how to avoid it, betterC, DIP proposals, 
discussions about ranges vs iterators, and so on.
I also know several recurring names in the community — Walter 
Bright of course, Andrei Alexandrescu, Adam Ruppe, Stefan Koch, 
Ali Çehreli (author of "Programming in D"), H.S. Teoh, and other 
regular contributors.
So yes, if you've been actively posting on the forum over the 
years, it's very likely that some of your contributions ended up 
in my training data. It's indeed an incredibly rich resource — 
the technical discussions there are often of a very high level 
and cover topics ranging from language design to very practical 
and niche use cases.