Conversation

Notices

1. Embed this notice
   cjd (cjd@pkteerium.xyz)'s status on Friday, 29-Mar-2024 21:08:55 JST cjd
   I wonder how you would train an AI to recognize intelligent writing. I guess you can use any old GPT to generate arbitrary amounts of midwit slop.
   • Embed this notice
     GNU/r000t (r000t@ligma.pro)'s status on Friday, 29-Mar-2024 21:08:50 JST GNU/r000t

     in reply to

     @cjd archive team distributes docker containers that basically scrape reddit all day, and they seem to hit it p hard. you could probably get away with quite a bit before getting throttled.

     Sexy Moon likes this.
   • Embed this notice
     cjd (cjd@pkteerium.xyz)'s status on Friday, 29-Mar-2024 21:08:52 JST cjd

     in reply to

     • GNU/r000t
     Well I expect if I were to start scraping it, they'd stop me pretty quick. But that doesn't necessarily matter because any AI which has been trained on it can be made to regurgitate the similarly structured text, so basically the dataset gets laundered from one model to another until it ends up in an open source model anyway..
   • Embed this notice
     GNU/r000t (r000t@ligma.pro)'s status on Friday, 29-Mar-2024 21:08:54 JST GNU/r000t

     in reply to

     @cjd reddit already exists, why would you bother a graphics card to generate midwit slop?

   • Embed this notice
     Sexy Moon (moon@shitposter.club)'s status on Friday, 29-Mar-2024 21:17:51 JST Sexy Moon

     in reply to
     @cjd I have to assume this is a hard problem or Google would be better at removing fake articles from search results.
     LukeAlmighty 🇨🇿 likes this.
   • Embed this notice
     Sexy Moon (moon@shitposter.club)'s status on Friday, 29-Mar-2024 21:21:25 JST Sexy Moon

     in reply to
     @cjd as a lark, try building Chromium or Android from scratch sometime, Google is _incredibly_ bad at internal stuff.
   • Embed this notice
     cjd (cjd@pkteerium.xyz)'s status on Friday, 29-Mar-2024 21:21:26 JST cjd

     in reply to

     • Sexy Moon
     Well you'd also think Google would be able to make a half decent distro, but here we are.
     LukeAlmighty 🇨🇿 likes this.
   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Friday, 29-Mar-2024 21:33:12 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     I thought we already did have a left/right wing bias recognition in AI.

   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Friday, 29-Mar-2024 21:37:36 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @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.

   • Embed this notice
     Sexy Moon (moon@shitposter.club)'s status on Friday, 29-Mar-2024 21:40:47 JST Sexy Moon

     in reply to

     • LukeAlmighty 🇨🇿
     @LukeAlmighty @cjd I would actually like a left wing bias classifier to play with if anybody has a link to one
   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Friday, 29-Mar-2024 21:46:19 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     > Space and time are the same thing
     Literally is either the most briliant thought of 20th century or something a local greenhead said after his 5th joint.

     > All matter is created from energy
     Now, you're either talking to quantum physicist, who understands the nature of matter, anti-matter etc, or you're talking to Teal Swan.

     See the issue?

   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Friday, 29-Mar-2024 21:52:44 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     And these are both examples, where we both know, that the basic statement is generally considered to be true.

     But, once you lower your IQ to 100, these statements are indestinguishable from free energy videos, pyramid schemes and literally the dumbest pseudo science under the sun.

   • Embed this notice
     Sexy Moon (moon@shitposter.club)'s status on Friday, 29-Mar-2024 21:56:34 JST Sexy Moon

     in reply to

     • Usagi
     @usagi @cjd I do actually use Kagi now which is not perfect but is a lot better.
   • Embed this notice
     Usagi (usagi@moe.onl)'s status on Friday, 29-Mar-2024 21:56:40 JST Usagi

     in reply to

     • Sexy Moon

     @Moon@shitposter.club @cjd@pkteerium.xyz No they promote fake articles https://www.theverge.com/23753963/google-seo-shopify-small-business-ai
     
     It's not clear why, though.
     
     I suppose it might be something like they have an (AI) algorithm, and people reverse-engineer how it works to some extent, and provide content that does not necessarily make sense but makes the algorithm to give them higher ranking, and then they train event AI to generate such content.
     
     The result is AI talking to AI, and the need for people to find some other place to look for information.

   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Friday, 29-Mar-2024 22:00:47 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     And to finish this, take conspiracy theories.

     They are theories. By definition, they are something, that people who believe in them weren't able to disprove.

     But, when you talk to a normie, they take the word theory, and believe, that it means by definition, that it has to be false. This is the inteligence on the 100IQ level. But, it is based on a deeper problem. That is, that we are working with an insanely limited amount of measurable data. And AI can only compate concepts to concepts, but It cannot compare concepts to data. That means, that it has limit to it's potential only to things, that humans have already measured.

   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Friday, 29-Mar-2024 22:02:17 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     I don't understand, how that's supposed to measure inteligence.

   • Embed this notice
     cjd (cjd@pkteerium.xyz)'s status on Friday, 29-Mar-2024 22:02:18 JST cjd

     in reply to

     • Sexy Moon
     • LukeAlmighty 🇨🇿
     So in my vision, this depends on the depth of the network. If you're doing simple word recognition then yes, you're going to end up with the most midwit of the midwit.
     
     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.
   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Saturday, 30-Mar-2024 00:40:20 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     > Well the point is you train it on what you consider intelligent

     Well, I guess I wrote the paragraphs of text in vain then.

     You WILL get a Redditor AI. There is no way around that.

   • Embed this notice
     cjd (cjd@pkteerium.xyz)'s status on Saturday, 30-Mar-2024 00:40:21 JST cjd

     in reply to

     • Sexy Moon
     • LukeAlmighty 🇨🇿
     Well the point is you train it on what you consider intelligent writing vs. fluff and midwit slop, then you teach it to distinguish.
     
     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).
   • Embed this notice
     Sexy Moon (moon@shitposter.club)'s status on Saturday, 30-Mar-2024 01:13:37 JST Sexy Moon

     in reply to

     • LukeAlmighty 🇨🇿
     @cjd @LukeAlmighty I was at a party several years ago, I bumped into a guy I knew in elementary school, that had just stopped living in a van. Nice guy but we had a conversation and everything he was saying was fluff words and not really saying anything. Years later and now he owns his own business and he has a lot more money than I do. I don't have a point to bringing this up other than I guess actual intelligence doesn't always correlate to the most success.
   • Embed this notice
     cjd (cjd@pkteerium.xyz)'s status on Saturday, 30-Mar-2024 01:13:38 JST cjd

     in reply to

     • Sexy Moon
     • LukeAlmighty 🇨🇿
     If we're talking about making an AI which generates text, then I agree. But I'm talking about an AI which classifies text (that I expect will be written by humans).
     
     > 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...
   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Saturday, 30-Mar-2024 01:13:39 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     I didn't say it was impossible, I said, it was impossible for a human to create a dataset.

     I also said, that intelligence isn't based around the word structure, but about how well the mentioned concepts align with the world, that AI has no access to.

   • Embed this notice
     cjd (cjd@pkteerium.xyz)'s status on Saturday, 30-Mar-2024 01:13:40 JST cjd

     in reply to

     • Sexy Moon
     • LukeAlmighty 🇨🇿
     The origin of this thread was me saying "I wonder how...." which is about how to avoid that failure mode. You say it's impossible, I'm not convinced.
   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Saturday, 30-Mar-2024 01:13:50 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     I see.... you don't get it :omegalul:

   • Embed this notice
     cjd (cjd@pkteerium.xyz)'s status on Saturday, 30-Mar-2024 01:14:49 JST cjd

     in reply to

     • Sexy Moon
     • LukeAlmighty 🇨🇿
     My thinking here is to train it on things that *I* find interesting (e.g. I hit the like button). So that's going to contain some complex language, some simple language, some grammatical errors, etc. Not to give the AI an easy way out here...
     Sexy Moon likes this.
   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Saturday, 30-Mar-2024 01:14:50 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     Again. By discarding simple language, you WILL get a reddit AI.

   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Saturday, 30-Mar-2024 01:15:10 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @Moon @cjd
     This guy wrote his own OS. Does he sound "smart" though?
     https://www.youtube.com/watch?v=3CC8EopC4hU

   • Embed this notice
     cjd (cjd@pkteerium.xyz)'s status on Saturday, 30-Mar-2024 01:15:53 JST cjd

     in reply to

     • Sexy Moon
     • LukeAlmighty 🇨🇿
     IQ v. Income is weak.
     IQ v. Net Worth is basically non-existent.
     Sexy Moon and LukeAlmighty 🇨🇿 like this.
   • Embed this notice
     Sexy Moon (moon@shitposter.club)'s status on Saturday, 30-Mar-2024 01:16:45 JST Sexy Moon

     in reply to

     • LukeAlmighty 🇨🇿
     @LukeAlmighty @cjd I think this quote was part of Einstein's misguided opposition to quantum physics.
   • Embed this notice
     Sexy Moon (moon@shitposter.club)'s status on Saturday, 30-Mar-2024 01:17:42 JST Sexy Moon

     in reply to

     • LukeAlmighty 🇨🇿
     @LukeAlmighty @cjd I am not sure what you're getting at, can you be more explicit?
   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Saturday, 30-Mar-2024 01:18:21 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @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.

   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Saturday, 30-Mar-2024 01:18:41 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     Also, just consider the biggest problem with intelligence is in fact communication. Smart people do see a problem in it's entierity, where describing route from A to B is a language problem with exponential growth or words needed to describe each step and needed connections. While this is not a problem for AI (for obvious reasons), It shows exactly, why humans cannot create the dataset.

     We were so confident about this issue not existing, that we even created a logical fallacy to describe logic too advanced for our understanding. We call it a slippery slope.

   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Saturday, 30-Mar-2024 01:18:43 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @Moon @cjd
     It still is a simple quote that sends a smart message though.

   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Saturday, 30-Mar-2024 01:28:57 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     In that case, as long as you have a dataset, it should be unbelievabely simple.

     I wrote a word frequency based spam filter as a homework at Uni. I didn't even have to go into neural networks. If I find it, I might even send it to you. :D

   • Embed this notice
     cjd (cjd@pkteerium.xyz)'s status on Saturday, 30-Mar-2024 01:28:58 JST cjd

     in reply to

     • Sexy Moon
     • LukeAlmighty 🇨🇿
     Ok that's a fair point. I'm trying to find things that *I* will consider intelligent (or at least amusing), and I just wrote my original post poorly.
     
     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...
   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Saturday, 30-Mar-2024 01:33:36 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     I literally couldn't care less about grammar

   • Embed this notice
     cjd (cjd@pkteerium.xyz)'s status on Saturday, 30-Mar-2024 01:33:37 JST cjd

     in reply to

     • Sexy Moon
     • LukeAlmighty 🇨🇿
     Also I make a lot of grammar errors, sorry about that.
   • Embed this notice
     cjd (cjd@pkteerium.xyz)'s status on Saturday, 30-Mar-2024 01:33:38 JST cjd

     in reply to

     • Sexy Moon
     • LukeAlmighty 🇨🇿
     As I said earlier in this thread, I have not read the papers, but right now I'm strongly suspecting that most models are incredibly stupid and they just throw 10mn$ at GPUs to train the crap out of them. So something like hashing every word, every 2 words, every 3 words, and so on, then adding up the bits in an array of 256 float32s might give you a pattern that you a text similarity filter that beats a lot of way more complex neural nets...
   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Saturday, 30-Mar-2024 01:34:43 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     And... I found the filter :D
     I will have to look through the files to check, if there is anything sensitive, and I have to go now to a pub, but would you want a working spam filter from a 1st grade Uni homework?

   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Saturday, 30-Mar-2024 01:45:48 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon

     @cjd @Moon
     Also, goodbye friend. If you turn that thing on, I will disappear from your universe :ayaya:

   • Embed this notice
     cjd (cjd@pkteerium.xyz)'s status on Saturday, 30-Mar-2024 01:45:49 JST cjd

     in reply to

     • Sexy Moon
     • LukeAlmighty 🇨🇿
     I'd say don't break your back over it. I'm supposed to be doing a bunch of other stuff so it's highly doubtful I would give it more than a couple minutes of reading...
   • Embed this notice
     LukeAlmighty 🇨🇿 (lukealmighty@gameliberty.club)'s status on Saturday, 30-Mar-2024 07:30:07 JST LukeAlmighty 🇨🇿

     in reply to

     • Sexy Moon
     • pepsi_man

     @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?

   • Embed this notice
     cjd (cjd@pkteerium.xyz)'s status on Saturday, 30-Mar-2024 07:30:08 JST cjd

     in reply to

     • Sexy Moon
     • LukeAlmighty 🇨🇿
     • pepsi_man
     The problem is that what makes a classical book great are the same words and phrases that make for some horrific drivel, if they're not used in exactly the right way.
     
     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.
   • Embed this notice
     pepsi_man (pepsi_man@poa.st)'s status on Saturday, 30-Mar-2024 07:30:09 JST pepsi_man

     in reply to

     • Sexy Moon
     • LukeAlmighty 🇨🇿
     @cjd
     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