One thing I'd be extremely interested in is a system that can improve itself at specific tasks or types of task. For example, many of the available LLMs don't excel at the type of writing I like, and rather than having to curate thousands of examples of a very particular and extremely rare niche, it would be nice to be able to describe what I value about just a few examples, and let the model improve on its own until it reaches a competitive level.

I'm talking about a hypothetical future model that is somehow able to model the type of curiosity and drive to genuinely improve of some real people. I don't think we've achieved models being able to innately tell whether they've improved the same way a person can. I can tell whether I've improved at art, at writing, at production, etc. without needing to be graded by a higher authority. I'd still like to check in and let the model know if it's been getting it right, but I really hope this sort of mostly unsupervised learning happens within my lifetime, because it'd be really cool.

Watch out for misuse of technical jargon. Big red flag. Remember those articles on arxiv claiming to solve reimann that no one can understand because of the disconnected writing style full of technical terms?