Page 1 of 1

LLM "support"

Posted: September 12th, 2026, 6:56 pm
by LuveelVoom
The fix for the scraper incident a few months ago has caused an unintended side effect: LLMs can no longer access ConwayLife and associated components (LifeWiki, the forums, possibly the book). While I am not a particularly big fan of LLMs, this content is uncopyrighted and free for any human to view; I feel like LLMs should be able to view it as well, so as to increase the utility of LLMs for completing CA-related tasks. I propose that we host a scrape of the forums and wiki on a separate site without cloudflare protection and update it every few weeks or so.

The main issue I see arising here is attachments. While small RLE files are easy to attach, images and large files are very difficult to store properly. I don't think this is a big issue, however, since most of the large files would not be of much help to LLMs.

Edit: can those who answer with one of the central 3 please say what they think should be changed/done instead

Re: LLM "support"

Posted: September 12th, 2026, 11:49 pm
by Disaster16439
For novice computer-ers like me, having an LLM that doesn’t talk nonsense about CA every 3 seconds is very helpful, I have to say, for installing programs/coding. It also really bugs me, to be honest, when LLMs talk nonsense. :D

Of course, there are downsides, but I think the attachment issue can be solved by just not includign them. It’s not like an LLM needs the EXACT rle of say, unidimensional spaceship 1.

Re: LLM "support"

Posted: September 13th, 2026, 4:17 am
by Gustone
LuveelVoom wrote: September 12th, 2026, 6:56 pmhost
I think that if it was possible to host something that could handle LLM scrapers, it would be equally possible to make this website able to handle LLM scrapers.

Re: LLM "support"

Posted: September 13th, 2026, 11:57 am
by LuveelVoom
Gustone wrote: September 13th, 2026, 4:17 am
LuveelVoom wrote: September 12th, 2026, 6:56 pmhost
I think that if it was possible to host something that could handle LLM scrapers, it would be equally possible to make this website able to handle LLM scrapers.
The main reason LLM scrapers bring down websites is because they follow every hyperlink, which typically leads to them getting stuck in loops. A webpage with only one page would have much lower load.