I use LLMs in a way that reduces social anxiety from my autism, I give it details of a strange social interaction that I could not parse on my own and ask if I should worry about it, or if I should make any kind of amends or inquiries, or if I'm over thinking something and leave it alone.
I use LLMs to bounce my own ideas off of that I'm not comfortable bouncing off someone I know IRL.
I use LLM's to role play. (all kinds)
I use LLM's to find things that I can't find via conventional research methods.
And you know what, my perspective on using it for "productive/generative" usage is nuanced. I get why artists and writers are upset, however there is nothing magical about human's and their artistic abilities and in terms of material economic impacts automation of various kinds has screwed working people in the past and generally I've seen a lot less push back.
I do think that generated images and writing is pretty bland and near worthless though without a ton of human done work atm anyway. Like, sure I could generate a video of a cat dancing on a moving bus while a nuclear bomb is going off in the background or whatever wacky shit with a simple prompt but what exactly am I even going to do with that?
Highly directed AI content that includes a lot of human work tends to actually be pretty amazing IMO.
And even though all the outrage pertains to intellectual work, this technology is going to likely result in a lot of blue collar work being automated via "embodied" neural network AI's. In fact, it may be that it was needed for this kind of automation to really take off at all. Its not just white collar work. We aren't just automating slop content and corporate purposed art. The day is coming when stuff like laundry, factory/warehouse work, and kitchen work, etc. is also all being done by robots.