TheDuke wrote: Thu Feb 06, 2025 6:39 am
Jonesy wrote: Thu Feb 06, 2025 4:37 am
So, do you not believe in the possibility of superintelligence and singularity? How can you differentiate between the chemical reactions and electric pulses in the human brain and the circuitry and algorithms of a computer? I’m not arguing the fundamental duality of human nature compared to a computer/algorithm.
I just think of it like this: the Internet is the current hub of all human knowledge— based on reality or our own flawed understandings. Integrating AI to the Internet with extremely fast processing speeds mimics the human brain. I would assume fast processing speeds and a vast information base correlate to higher AI intelligence; this along with machine learning and advanced analytical capabilities, how can this not eventually lead to super intelligence? Comparing AI to bots seems way out of touch at this point!
So what is the danger between thinking it is intelligent or based on algorithms if our interactions with it is comparable and almost indistinguishable with another human? Seems dangerous whether intelligent or not! AI is already surpassing average human intelligence. I can envision androids being built that mimic the human body and can use senses mimicking ours. Many years away, but who knows. I’ve heard some scary things that some AI, before being deleted, would try and save itself by hiding code for self preservation. Stuff like that! We also have to account for the eventual breakthroughs of nano technology—another game changer.
I'm not sure where even to start. first. the INTERNET has NO KNOWLEDGE at all........... nothing. perhaps facts, mingled with lies and falsitudes of all sorts. but "intelligence", no. Intelligence is a kin to "wisdom". It requires more than facts or points or issues, or statements or assertions, or data. Intelligence uses facts.
I often see issue with discussing "truth". Truth is not a statement of facts...... the scriptural definition of truth is "the knowledge of.............."". Knowledge is the abilit to create and or use facts/ or "factoid" or what ever you wish to call all the "data" or "information" points (usually AI assertions technically) for a purpose.
I don't believe in a computer being able to create knowledge. It is able to dig through "factoids" (without any basis of judgement) and find corrolary factoids, And true AI's purpose is to create new "factoids" or assertions by combining existing ones. like a = b, and b = c and we have a rule (AI inference rule) that if any two things are equal so must be the third, so it creates a new factoid or assertion (pick your technical term) and says a = c. Now, in reality there is never (not often anyway) anything as clear as this, so there are shades of gray. This is called "fuzzy" and requires "fuzzy logic". If I think I have a, and I think a equals b, then may assert (infer if you like) that a equals c. But, there is an added fuzziness or probability or tenuous nature. Sometime other factoids and increase the probability of your new "term" or factoid, or assertion. but eventually, it becomes too fuzzy, several iterations into the logic cycle. It appears that it is learning, and making new facts. And on the surface it actually creates such things. But then things go awry. Why, because the data is imperfect, as are the inferencing rules, and the inference engine is cyclical (not running to total optimization).
So, you start creating factoids (as facts) that are more and more fuzzy or tenuous, until they are just wrong. The head of the AI group (supposedly open AI, bah) call is "halucenations" and says they'll go away....... they can not go away but get carried along like rocks in a glacier, and deeper embedded.
There is no intelligence, artificial or otherwise. Worse, any on here see a natural language interface, that talks nice and makes up crap and thinks that is intelligent, cause it sounds nice, but a parakeet can talk back to you, quite clearly. It is the same with images and image recognition. There is no intelligence, just looking for repeatable features. Many errors creeping in.
So, no, I don't accept some algorithm will ever be really intelligent. It may be able to play chess, guess the stockmarket (and drive it if enough people agree) and even tell people what to do and think (but so does Taylor Swift for that matter).
an algorithm is only as good as the programmer and the data base and the ability to diagnose and repair it. BTW with learning algorithms, there is no way to ever perform any good quality control as the algorithm is different than when you start the dianosis.
People need to really read what AI is, not what some liberal or non-scientific editor says it is. And there is a place for it. I mean we have been using these techniques since the first search engines started indexing the HTML and other data files. Some of us have been working with the techniques since the origination of SQL and such searches. very good for sifting lots of data and looking for specific things.
My wife loves her AI interface to help her write letters. She is a Spanish speaker and even a marginal english language bot is nice and productive. and more patient than me. Usually it is correct, but sometimes it says stupid things. I mean anyone that has had a text or email message changed by a language assistant knows what I mean.
the problem with the AI sold today, is people use it, mostly for superficial things, the accept it. Sort of like we all were in our TBM days and accepted things that seemed correct and consistent but now we find were altered and changed many decades ago.......... truth burried beneath heaps of repeating and made up consistency checking.
but there is way too much information and data to ever sift through and the thing about the AI ;push is the blending of original information or factoids and those added.
Like here we have a plastic straw ban....... some research paper said how it was killing dolphins, so goody-two-shoes got them banned. Now we find the paper was written by a mom for a kid's HS science project and she made up the facts (factoids) of deaths of dolphins, etc... .it got repeated in some college papers, then in science journals, and now she says it was made up and all those science journal call her a liar............... yet the pedigree can be traced. but we cannot purge the follow on lies or new factoids created. as they are fuzzy and we can only search certain terms. plus TPTB like banning straws anyway. BTW I never use straws so I don't care, but it is a small example or things being said and becoming "facts". Could do a long tutorial on COVID and shots to show how many false factoids exist, or climate change, could go on......... those are what the AI uses. As do people with no intelligence of their own....... and there are quite a few it turns out, IMO.
269 km SE of Pondaguitan, Philippines 
April 22, 2025 — Earth Day 
04:17:13 UTC-06 
Magnitude 6.2 
Depth: 117.8 km
117.8 km deep — a whisper not from the crust, but from the deep blue codex.
Interpretation for the Archive:
Earth quaked. 
The timeline moved.
Society fears the thinkers who also speak. 
A tale as old as time, yet recent as the now. 
**Disconnected.** 
; much stranger yet!