you are viewing a single comment's thread
view the rest of the comments
[–] 9 points 5 months ago

I’ve been saying this for a while to people. I think the long term use case for LLMs is the semantic human interface device.

Siri,Alexa, even google home?(whatever they called it), they all swung and missed at this. However evening able to provide commands unclearly to a computer and get the intended result would be a huge win.

I know the big llm inference can do a lot more but the cost is high for systems with that ability to reason however small, lightweight llms are actually very good for command and control.

This where my current homeland projects are focused.

  • source
  • parent