There are all sorts of interesting demos of semi-autonomous agents built with LLMs, and impressive intent-recognition capabilities of these models. Curious if there are repeated tasks in your life that you wish you could hand off to an AI system to get off your plate? (Ideally things you think should be doable in the near future with current state of the art)
Laundry and washing dishes.<p>Stacking plates, vacuuming, and folding clothes.<p>Scrape a web UI and infer an API.<p>From a gameplay video, create the game.<p>From a storyboard, create the video.<p>From a screenplay, create the comic.<p>From a sketch, generate a 3d model.<p>Prove the <i>form</i> of an argument is valid, including proof steps. (<i>Not</i> whether premises are actually true, etc.)