Sounds pie in the sky. Give examples ?AI to independently train on the XenForo database and generate answers unique to this community
By whome? And on which foundation? It might be the case that you would like to see an AI integration - but that is very far from "it is recommended". There are some people that would like to see AI in XenForo while for others it would be a reason to look for a different forum software.It is recommended to integrate DeepSeek AI functions into the search box.
It is strongly recommended to add an algorithm module and set up a “Discovery” tab at the top. Upon entering this tab, it should display all the topics of interest to the user based on the calculations derived from their browsing data.
This reads like an AI generated answer, given the combination of tone, lacking substance, dramafulness and complete lack of understandable and useful content.View attachment 324684
The ima app mentioned above can use AI (such as DeepSeek) to generate exclusive answers for users' private databases, independent of the entire network database. Try it or Check the functions of this app on Xiaohongshu.
Algorithms and AI-powered apps occupy most of users' time, and ordinary communities only survive in the cracks.
I hope XenForo gets better and better and remains technically advanced. But frankly, community software without algorithm recommendations and AI is really outdated.
Below are my previous recommendations based on algorithmic recommendations.
It most certainly isn't. IMHO most users who are happy with recommendations and AI answers will not visit forums any longer sooner or later - and those who still want to use forums don't want AI assistence, they want to talk to humans.But frankly, community software without algorithm recommendations and AI is really outdated.
Algorithm-based content recommendations and AI-generated responses are not divorced from human interaction. Let's wait two to three years and see what features the major mainstream community software will have.It most certainly isn't. IMHO most users who are happy with recommendations and AI answers will not visit forums any longer sooner or later - and those who still want to use forums don't want AI assistence, they want to talk to humans.
This is the key selling point for forums, so adding user-facing AI features might in fact be counter productive.
As soon as human forum or social media users fear that AI will answer their questions, they will return to more places where real people can answer their questions.It is recommended to integrate DeepSeek AI functions into the search box.
Using AI to independently train on the XenForo database and generate answers unique to this community.
You can refer to the app "ima".
Think about the diffences between Tiktok and a forum. If you prefer TikTok - why don't you become a creator and consumer there?Think about why algorithm-based social media apps like TikTok take up the vast majority of users' time.
BTW: DeepSeek is a Chinese company. South Korea, Italy, Taiwan and Australia have already taken action against them due to privacy concerns. Germany is about to start an investigation to ban it from the app stores. In the US, employees of several institutions like i.e. NASA and minsetry of defense are not allowed to use it. So problaby far from a good recommendation. For privacy reasons and as probably the use of XF would possibly fall under regulation in these countries if DeepSeek was integrated. Integrating DeepSeek into XF would thus be suicidal.It is recommended to integrate DeepSeek AI functions into the search box.
I don't actually have an issue with this one and, done right, I would implement it.Algorithm-based content recommendations
Depends. I.e. the biggest version of Metas Llama (LLaMA 3.1 405B) requires 750GB to 2TB of disc space, the whole repository is 4,9TB. However, there are smaller versions. You should also not ignore the other hardware requirements, namely VRAM:was just wondering how much storage space and resources would a locally hosted open source llm require to just locally process the content of the forum and provide a Q&A interface for responding to queries right on the forum.
We use essential cookies to make this site work, and optional cookies to enhance your experience.