@cjd @Moon
Obvious jokes aside, the problem is, that you cannot create a dataset by humans, since humans are incapable of making this distinction themselves.
The entire concept of schizophrenia and inteligence being 2 sides of the same coin does apply here 10fold. Because, briliant people do see paterns that you cannot visualize, that means, that you cannot know, if they actually are smart, or if they are bullshitters.
This is, why most atempts at doing this end up just with recognizing, how niche words you use, since the niche words are needed to make a scientific article. But, you immediately turn to social science loons, who cannot form a single sentence without going full systemic prejudice against margenalized methaphors for cheese.
But, and this is a simple implementation: Suppose you use a text classifier on the individual sub-phrases, then for each one of those, you output neural layer snapshot represented as an image, then you take the images making up a sentence, and you feed them through a net to pattern recognize similar sentences and again you're outputting a neural snapshot as an image.
At each level of this, you can train using 2 similar phrases and one different. The reward function is based on the neural image of the similar phrases being more similar (XOR of the pixels is less) than the different one.
Feed those images back in, this time per-paragraph, and you should have a form of paragraph level classification. Then you feed that output into a network which which classifies text into a score and you train on things you find worthwhile.
BTW humans have a way of signaling and detecting intelligence - that is through humor. It's like the first man-made proof-of-work: It requires more brain cells to be funny (prove) than it does to laugh (validate).
> isn't based around the word structure
Well, yes and no. REALLY stupid text has an identifiable structure. Midwit text looks smarter than it is. What I'm looking for is how to make the model deep enough to identify quality fedi banter.
Of course midwit diarrhea is a moving target w/ Goodhart's law, especially if people are start training GPTs against my "quality posts" classifier...
@cjd @Moon
If you want to make an AI to learn what interests YOU, then even the dumbest "find words I like" system will do the job. As long as the text contains "linux, boot, freeware, software, hardware", It's great.
If it contains "republican, democrat, trump, fuck, Nigger" It's BAD.....
But you changed the goal entierly now. Your original post was about finding intelligence.
Can this AI read books? If so, then have the AI read certain "men of letters" (blank slate AI).
Perhaps beforehand, have it learn to distinguish verbose writing (too many adjectives, too many adverbs, and buzzwords)
A bank of buzzwords can be maintained quite easily, and have any pop-culture article containing these words be flagged as stupid or low-brow. (This is giving the AI some agency.)
Anyway. My wheelhouse is mathematics and Tech. Comm. and not software engineering, so throw rocks if you like.
@LukeAlmighty @Moon
Training an AI on quality writing is interesting, but tweets and fedi posts are perhaps better for training because they're short and so it doesn't take reading over pages and pages of long words to determine whether you're dealing with genius or reddit-tier poop.
@cjd @pepsi_man @Moon
I am sorry, but I thought you are very educated when it comes to the IT.
So, why would you believe, that less information is good for learning?
But the thing is, I can't think of any particular set of words that communicate what I would find interesting - it's like trying to word-filter for what you'll find funny. What do you do? Filter for "knock knock"? Maybe if you're 5.
Most political takes are boring and repetitive. Most Science is horrifying midwittery. Most blockchain takes are spam and get-rich-quick. Most conspiracy takes are aliens and flat earth bullshit. BUT, there's 1% in each category which is a flash of brilliance (IMO) and I'd really like to try to filter for it...