![](https://programming.dev/pictrs/image/87c50ac6-70dd-4718-b933-85cd40f15a7e.png)
![](https://beehaw.org/pictrs/image/c0e83ceb-b7e5-41b4-9b76-bfd152dd8d00.png)
'tis how LLM chatbots work. LLMs by design are autocomplete on steroids, so they can predict what the next word should be in a sequence. If you give it something like:
Here is a conversation between the user and a chatbot. <insert description of chatbot>
<insert chat history here>
User: <insert user message here>
Chatbot:
Then it’ll fill in a sentence to best fit that prompt, much like a creative writing exercise
It was initially intended to be a video stream handler, but they had concerns with audio syncing. They figured they might as well also handle audio in one cohesive AV server instead