@Ali @Fox @antoviaque @jordan @pooja @kaustav
This is just an FYI. I’d like to turn all your 2026 Open edX Conference talks into blog posts. I’ll create tickets and share them here 
Well done to everyone this week!
PS: @Agrendalath I’ll take your speaker notes and work on ours 
@Ali @Fox @antoviaque @jordan @pooja @kaustav
I’ve created a parent task with subtasks. Please let me know if the timing doesn’t work for you directly on the ticket. You’re welcome to work ahead on it.
@cassie Could we find a more efficient way to convert them, than spending 10h on each? That is a lot of hours invested. Imho one person could work with an LLM to produce a good base for all the articles based on the presentation video transcripts, and then have the individual authors only review & edit the articles before publication.
@antoviaque I do agree we should use LLMs to make this process more efficient. The original 10 hour estimate was more based on our usual end-to-end blog workflow (writing, reviews, editing, SEO optimisation, images, formatting, publishing etc.), rather than this specific “conference talk to blog post” scenario where we already have a lot of the source material.
Since we already have talks, scripts, slides, and transcripts, I think we can probably reduce the effort quite a bit, likely closer to 5 hours per post with LLM assistance.
That said, there’s still a fair amount of editorial and publishing work involved beyond generating a first draft. For example:
I think your proposed workflow could definitely work if someone wants to own the coordination/LLM drafting process. I don’t currently have bandwidth to generate and coordinate all the article drafts on top of my existing client work. Any takers?
My suggestion was more to distribute workload:
That way the workload is shared, posts refinement can happen in parallel, and we’re more likely to get everything published while the conference is still fresh 
I wasn’t really pinged here, but I did have a lightning talk and I do have the video of that since I recorded it and sent it for the conference. I also have the transcript, and I first did a rough write-up to get my ideas in writing so I think at least in my case I wouldn’t need anywhere close to 10 hrs to get a blog post out of it, closer to 2-3hrs at most.
@cassie I think the LLM prompting is something to do once for all video transcripts at once (when the videos are published), there is very little gain from having to redo it for each talk separately imho. Maybe just provide the prompt & raw transcript to the reviewers if they want to tune it further for their talk, but imho it should be possible to get the LLMs to produce almost publication-ready work for all the blog posts at once. Most of the work was done in the talks themselves.
And five hours of editing afterwards also sounds like a lot to me. When we discussed this initially, I had more in mind 1-2h of someone working with a LLM to produce the base blog posts, then 30min-1h of review/editing work for each afterwards. The time budget of 5-10h/talk is more the time needed to prepare a talk from scratch…
Noted. I’ve kept a ticket for everyone though. As in some cases I’ve requested the original script / speaker notes which can be fed into an LLM. That way I don’t have to wait for them to release their videos as that has taken a bit of time in the past. And I want to get some content on the site soon as we’re breaking ground with our keyword strategy.
I’ll also ask everyone to do a quick read over their own to ensure there are no errors (especially for the technical talks).