What makes AI model better with Hollywood?

[lazi]

I had a successful interaction with Gemini AI and it accepts corrections from me.

have noted these corrections and will remember to apply the correct Hollywood programming language syntax in our future interactions. This includes using <> for inequality, #TICKS for time units, BeginDoubleBuffer() and EndDoubleBuffer() for buffering, and passing colors as direct arguments in drawing functions.

You can always ask me to forget things or manage what I've saved for you on the Saved info page.

Will this affect your language model for other users?

No, these corrections will not affect the model for other users.
The information I've "remembered" is stored specifically within your personal user profile

So the question is: what should we do to improve such LLMs?
- more code snippets on forums? (which forums)
- books on Hollywood programming?
- youtube videos on Hollywood tutorials?

oh, maybe I should ask Gemini itself!

Any thoughts?

[plouf]

code snipets with good comments
also we can train one,

I was thiking about this, soem time, but time is pressure
however if you like we can assemble a team to train one to see results

afaik it must provide thousand of examples to start produce results