Ha, Whoops. If I could (without using AI) determine if content was likely to be AI, using this tool, I would, but I do not know how that would be possible. Research perhaps for another day, or maybe that would be worthy of DHF funding.
The irony being, a lot of the code for this tool was translated from my rough python psuedo code into JS using a few different LLMs.
JS and Python aren't all that different, but I am spending the day today fixing silly and stupid logic in a bid to make the tool run faster, so it can scale as the scope and quantity of content changes.
I am ever a delusional optimist about the future of hive being its human generated content.
It wasn't a complaint as I know filtering out AI generated posts is borderline impossible. I was just mentioning how stupid people are. I haven't checked the stats of the master of tasks, most likely I'd see all academic level posts as we know how they operate. Anyway, it's just for fun as hive should be about having some fun and not writing dissertations each day. I suppose just as with reputation (which we know some got theirs using bidbots), everything here has to be taken with a pinch of salt.
Absolutely - by pinch of salt, I think you might mean those enormous mining trucks that need a ladder and hard hat to climb.
I am a data driven individual, and I am also a creative individual. I remember when I discovered the Flesch Kincaid measure in primary school, I was fascinated by how writing can be broken down using analysis.
The response from many writers to this tool on HIVE has already been incredibly positive, and that alone has made the time I've spent on it so worth it.
I have another use case for it, in which it might be able to give a hint at how an individual user is engaging on the chain, or we might be able to input a list of user names (once I fix all the bugs) to see how, perhaps; a group of newly created accounts have "performed" in terms of generating engagement in response to their content, and what sort of content tends to generate success.
This is all in an ideal world, however.
In shorter terms, for projects who claim to onboard "quality" users to Hive, I want to have (another) version of this tool that can iterate through users and then allow people to make educated decisions about how they vote for witnesses and how they direct their DHF funding.
Big picture, longer term stuff, which I have only thought of in my rambling reply to your reply.
I remain ever-delusional and ambitious.
I do, however, value your input - what is a feature that you would like to see - be it in the existing tool, or future versions?
Exactly! Taking everything with face value is not the best thing and we know it. New users don't, but that's not so important right now. Veterans know how things work here.
I am also a data driven person as data is my life, I can't work without data, so I understand you perfectly.
I have an idea for you, but I'd rather send you the details privately as it's better. Actually I have more than one idea, so drafting a list and will contact you on Discord shortly 😎😂
No problem, I look forward to your message. I have the same username on discord. I lurk as invisible, so send a friend request, and send a message :)