Weekly AI Summary on meta.discourse.org
Overview
This week on the meta.discourse.org forum, discussions revolved around various AI-related topics, including a proposed moderation tool for formatting code using AI, the ongoing development of Discourse AI features, issues with AI summarization backfills, and the potential for uploading and discussing PDFs within the composer. Additionally, there were conversations about setting usage limits for AI, evaluating costs between different AI providers, handling AI-generated spam, and exploring the capabilities of AI bots and custom tools.
Interesting Topics
-
@merefield raised an idea for a moderation tool that would allow trusted users to format code blocks using AI, potentially improving readability and assisting new users who struggle with proper code formatting.
-
Discussions continued on the Discourse AI plugin, with @sam expressing interest in allowing users to upload large files and ask questions about the content using a persona-based approach.
-
@markschmucker encountered an issue where the AI summarization backfill process kept regenerating summaries for the same topic, even after a valid summary was already present. This led to a fix by @Roman_Rizzi to make the job more resilient.
-
@BrianC proposed a feature request to allow users to upload PDFs or text files directly in the composer and have the AI process and respond to questions about the content.
-
There was a discussion around setting per-group token and usage limits for AI features, with @sam clarifying that quotas are defined per group and applied per user, rather than being shared among users.
-
@oppman sought advice on evaluating the costs of using HuggingFace TGI versus OpenAI API endpoints for AI services, with @Falco providing insights on when running one’s own language model makes sense.
-
There was a bug report regarding posts and accounts not being restored when flags from Discourse AI spam detection were rejected, which @sam addressed with a fix.
-
Discussions took place around self-hosting embeddings for DiscourseAI, with @sam mentioning ongoing work to restructure the embedding configuration and plans to support multi-model embeddings.
-
@David_Ghost inquired about the ability of AI Triage to perform searches and avoid topics with similar titles based on creation dates.
-
@smithmarkl asked about the possibility of injecting custom prompts or manipulating embeddings in the AI Search feature to boost or separate staff responses from community responses.
Activity
-
@merefield proposed an idea for a moderation tool to format code blocks using AI, with @Jagster and @sam providing suggestions and feedback.
-
@sam expressed interest in allowing users to upload large files and ask questions about the content using a persona-based approach in the Discourse AI plugin.
-
@markschmucker reported an issue with AI summarization backfills where summaries were being regenerated for the same topic, even after a valid summary was present. @Roman_Rizzi provided a fix to make the job more resilient.
-
@BrianC made a feature request to allow users to upload PDFs or text files and have the AI process and respond to questions about the content.
-
@sam clarified that AI token and usage limits are defined per group and applied per user, rather than being shared among users.
-
@oppman sought advice on evaluating the costs of HuggingFace TGI versus OpenAI API endpoints for AI services, with @Falco providing insights.
-
@sam addressed a bug report regarding posts and accounts not being restored when flags from Discourse AI spam detection were rejected, with a fix.
-
@sam mentioned ongoing work to restructure the embedding configuration for DiscourseAI and plans to support multi-model embeddings.
-
@David_Ghost inquired about the ability of AI Triage to perform searches and avoid topics with similar titles based on creation dates, with @sam indicating that such “agent-like” behaviors are being considered.
-
@smithmarkl asked about the possibility of injecting custom prompts or manipulating embeddings in the AI Search feature to boost or separate staff responses from community responses, with @Saif suggesting making a feature request.
-
@c12gene reported an issue with the AI bot not being able to read summaries and detailed content, which was addressed by @MachineScholar through system prompt improvements.
-
@Ricardo_Kutz and @CocoQuark shared their experiences with AI-based spam on their forums, with @j127 and @princejamesoo providing suggestions for handling such spam.
-
@huynhthai824 inquired about the capabilities of Discourse AI in suggesting relevant topics and sending private messages to users based on their queries.
-
@BrianC asked about tying token limits to subscriptions and allowing more expensive models to be used for a fee, with @sam confirming that different quotas can be set for different user groups.
-
@MachineScholar inquired about the possibility of inserting API keys and OpenAI project details from the admin settings in custom tools for the AI bot.
-
@huynhthai824 asked about the ability of an AI chat bot to aggregate and respond to real estate queries with min/max filtering on a Discourse forum, particularly for unstructured data.
-
@oppman sought recommendations for self-hosting an open-source language model for DiscourseAI, with @Eric_Keller sharing their experience with self-signed certificates.
-
@markschmucker discussed how regenerating summaries works and suggested changing the interval for the backfill job to improve efficiency.
-
@joo asked about modifying the Riona-AI-Agent project to integrate with Discourse’s API.
Thanks for reading, and I’ll see you again next week! 