You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Today, course generation is one-shot. Once a course is created, users can continue the conversation in the classroom, but those follow-up explanations do not become part of the course itself.
This creates a gap: when a user wants to go deeper on a specific point, the system can answer in chat, but it cannot naturally extend the course with new pages based on the current context.
I want to discuss adding a dynamic classroom capability at runtime:
Users still generate the initial course the same way as today.
Dynamic mode is a course-level toggle that can be turned on or off, including after the course has already been generated.
When enabled, the director can do one more thing in addition to orchestrating classroom agents: it can decide to append a small number of new outlines.
New outlines are appended to the end of the course only, typically 1 to 3 pages.
Page content, actions, TTS, and media generation continue to use the existing pipeline. This does not introduce a second page-generation flow.
For storage semantics, my current preference is:
The toggle only controls whether future dynamic extension is allowed. It does not rewrite existing pages.
Dynamically generated pages are written back into the current course and keep the same course ID.
If we want to preserve the original version, that should be an explicit duplicate / fork action, not an automatic new course on every dynamic extension.
This would keep the scope focused around director decisions, outline extension, and wiring into the existing generation pipeline.
Questions:
Should we build this feature at all?
When should the director append new pages, and when should it just reply normally?
How should we prevent new pages from overlapping too much with existing ones?
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
-
Today, course generation is one-shot. Once a course is created, users can continue the conversation in the classroom, but those follow-up explanations do not become part of the course itself.
This creates a gap: when a user wants to go deeper on a specific point, the system can answer in chat, but it cannot naturally extend the course with new pages based on the current context.
I want to discuss adding a dynamic classroom capability at runtime:
For storage semantics, my current preference is:
This would keep the scope focused around director decisions, outline extension, and wiring into the existing generation pipeline.
Questions:
Beta Was this translation helpful? Give feedback.
All reactions