Is @Discourse trying to sell itself as a "data source" for LLM training?.. 👀

[rysiek@mstdn.social]

RE: https://mastodon.social/@Discourse/115890721735332522

Is @Discourse trying to sell itself as a "data source" for LLM training?.. 👀

This really sounds as if the value of forums and communities was largely as a resource for training LLMs:

> Community content strategies should focus on creating discussions that encourage depth, comparison, and resolution rather than fleeting engagement. This is a long-term investment in knowledge infrastructure and one that serves both human learners and AI systems.

Sounds very corporatey, AI-bubbley.

[jonah@mastodon.neat.computer]

@rysiek what they want is for LLMs to be built into Discourse, which use the Discourse forum as a web source for research instead of just performing generic Google/web services like a typical researching cloud LLM would.

[miklo@fosstodon.org]

@rysiek @Discourse
The foundations were laid quite some time ago:
https://blog.discourse.org/2023/04/introducing-discourse-ai/
https://blog.discourse.org/2025/10/discourse-mcp-is-here/

[oblomov@sociale.network]

@miklo @rysiek I like @Discourse as a software, but this is quickly going the same way as StackOverflow and I do not like that.