Am I on Mod Queue?
#11
(2025-10-13, 12:31 AM)Omar G. Wrote: I have moved your post to the "MyBB Project & Site Issues" forum, and I am replying for notification reasons and will probably reply later.

Thank you.

Might be this is the first instance of my starting to fully appreciate the scope of this AI topic. Of course, I've briefed myself on some aspects of this and have seen this-or-that written by all sorts of smart folks; but deep thinking on my own hasn't been a strong point at my end - yet. I just wish my brain was working a bit better to get a better handle on this AI. But have to start someplace, and right here may not be such a bad start.

For example, as I think about this, somebody must have done some research on just how far a human can push vocabulary usage/sentence structuring upon an AI program before the AI shuts the human down due to not comprehending what the human is writing?

And now I have just used the vocabulary "comprehend" when referring to an AI program. That seems weird because it seems to fit!
Reply
#12
Note that the conditions and values are only examples - in practice, administrators would customize them and maintain a private ruleset according to characteristics of desired and undesired content on their forums.
devilshakerz.com/pgp (DF3A 34D9 A627 42E5 BC6A 6750 1F2F B8AA 28FF E1BC) ▪ keybase.io/devilshakerz
Reply
#13
(2025-10-15, 06:58 PM)Devilshakerz Wrote: Note that the conditions and values are only examples - in practice, administrators would customize them and maintain a private ruleset according to characteristics of desired and undesired content on their forums.

I'll have to be excused for not having the time to properly begin doing the necessary studying to delve deeper into this, but when I saw that category "INCOHERENT" I assumed this was a program where a certain level of evolving data gathering was being done by that program.

In fact, I already had the following in a folder from some notes related to the Columbia AI Summit:

https://ai.columbia.edu/news/ai-moves-ad...telligence

And one paragraph:
Quote:The key difference is that embodied AI learns through experience and interaction, much like humans,” Haddadin said. “It builds models of the world through sensory feedback and real-world interaction rather than just analyzing statistical data. One could say it is a curious discoverer in its own right.

But your post, Devilshakerz, seems to indicate that anything incoherent in the Ad Rem extension is defined by the webmaster/admin?

Or maybe I'm misunderstanding?

Again, I haven't yet found the time to properly dig into this; like taking a thorough look at the Ad Rem code.

An aside: I had weird thoughts when I hit that anti-bot thingy when logging in and asking about things that can stay up in the air. I could get any of those vehicles up in the air and for longer than those birds could stay in flight. A really, really smart bot would answer by clicking every image shown. A really, really stupid old man like me might try to do the same thing, but I suspect that hCaptcha thingy would then throw up a clever message: "Heh, stupid old man, don't try to trick me. I was trained by Hercule Poirot!"
Reply
#14
So I haven't created a time slot to be able to dig into that Ad Rem extension, which caught my interest. BUT, I don't wish to let this thread go unattended for too long, as I will be digging into the Ad Rem thing within a reasonable time, I still hope.

I thought, though, a quick reference to what has me a tad nervous about one aspect of this AI stuff could be easily represented by something like the following:

   

It is that rough idea that if a machine can develop the ability to use logic, we humans could be in a bit of trouble in about 100 years. Or less.

Well, maybe the use of "if" is already outdated.

And, yes, I still think of any AI sort of thing as a "machine". Of course, humans can mimic machines and if anyone has done active duty, you know exactly what I mean. Maybe early law enforcement training also used to use that "You Are A Machine" line of yelling from a DI. I think the modern military might be trying to cut some of that out, but not sure.

But a "machine" learning .to mimic a "human" makes me nervous. A lot!
Reply
#15
@scealc, how familiar are you with large-language models? I get the sense from your comments that you're not particularly aware of them, nor that you can freely access them and get a sense of where cutting-edge AI is at these days directly, for yourself.
Reply
#16
(2025-10-28, 03:25 AM)Laird Wrote: @scealc, how familiar are you with large-language models? I get the sense from your comments that you're not particularly aware of them, nor that you can freely access them and get a sense of where cutting-edge AI is at these days directly, for yourself.

I am totally ignorant on/of many-many things in life; and in non-life. And "large-language models" is one of many things I am ignorant of. Worse, is I do not care about large-language models. And the reason is that it doesn't faze me in the least when I am giving consideration to a possibility that a "machine" can become dominant on this Earth. Please note the use of "possibility" and as that has so many letters it must be an example of large language when compared to its very tiny brother/sister "a" - - - are words without gender? As a machine might be, if it gains logic - - - without gender, that is. Which then takes us into gene building. Humans are trying to build genes, right? What if machines learn to do it better?

Well, maybe "trying to" is incorrect - - - might be some science type human has already done it.

By the way, Laird, I mean no disrespect; but I could care less where we presently are with regard to AI development. I am worried about what's going to be happening in one hundred years. And I am sure to not be with the living by then. Well, almost sure.

Actually, that helps me summarize the whole thing; I care about AI developing itself, without the aid of humans.
Reply
#17
Oh, OK. In your previous post you said that the idea of a machine learning to mimic a human made you "nervous", and that if a machine "can develop the ability to use logic" then humans would be in "a bit of trouble", so I thought you'd want to know that LLMs can already do both to a meaningful extent, but if you now say that you "could care less", then, fine.

Re "AI developing itself, without the aid of humans", we are probably a lot closer to that than you might imagine. On this forum alone we already have an example of LLMs being used to code non-trivial MyBB plugins from scratch. It's probably not long before they're tinkering with their own code.
Reply
#18
Okay, I was a bit careless with my vocabulary usage. Sorry about that.

What I should have made clear is I don't care about individual code projects that help the AI to develop, as I thought/think that LLM project is. I'm already convinced that humans have started down that road. Trying to digest and understand all the projects ongoing would have to be for some younger code builders. I'm in my 8th decade of life and probably am guilty of helping some folks start down this awful path way back when it might have started. Heck, I used to help Google. Then I turned against that project and did it right in front of management and engineers and they got really pissed off at me. Point is, I am pretty guilty myself.

And I am also not so sure when I started to really get nervous about this AI stuff. Not that long ago, as best I can remember. My trouble with Google and Microsoft (another group that got pissed off at me) - - - my trouble with them was they seemed to be wanting too much money. Of course, I can't define "too much" and then my own family was able to buy into Google right at the start because of my connections and I get nailed for being a hypocrite. Except, I really didn't know what was going on over there in the U.S. and my family; which I've been branded as being untruthful about.

BUT I didn't have fear of the AI projects in the early years. Maybe I was just stupid. No, not "maybe" - - - I was. But I think a person can walk out of the realm of stupid; but maybe not the realm of machine control. And "machine control" would be when those machines are smarter than us.

Funny too, I seem to remember that some doctor or science type informed us that humans don't use their full brain capacity. Well, now it might be a good time for one of us to do so. And then teach other humans how to do that.

By the way, this bit of information is very-very much appreciated:

Quote:Re "AI developing itself, without the aid of humans", we are probably a lot closer to that than you might imagine. On this forum alone we already have an example of LLMs being used to code non-trivial MyBB plugins from scratch. It's probably not long before they're tinkering with their own code.

That was exactly what I was going to try and learn with regard to the Ad Rem project. BUT I don't think I have enough years left to be able to do anything about individual projects, so I have to go for the bigger picture; brainwash some younger folks with old-school thinking.

Of course, "old-school" thinking got us that darn atom bomb thing and so younger folks could claim we old types already messed up this Earth program.

I wonder what an AI elementary school teacher would be teaching? First there were these things called dinosaurs; then there was this thing called a human. Don't forget class, the dinosaurs were much smarter. The dinosaurs didn't build AI. Now we run this planet. And we don't need oxygen, so we are going out to the rest of this solar system.
Reply
#19
Just for clarity: LLMs are not an individual project; they're a class of AI technology, typically based on generative pre-trained transformers. There are many individual LLM projects: ChatGPT was the first one to be made public and become popular; now there are Gemini, Claude, DeepSeek, Grok, and many others.

I mentioned them in the first place because, again, most of them are freely accessible, and if you were to access any of them for yourself (click on any of the above links) and explore them in a conversation, you'd discover that your fears of what might come to be 100 years from now have already been realised (mostly, at a minimum).

I'm not saying that those fears are unfounded. There is still a serious risk if this technology is given too much power and control. So far, it hasn't been, and agent-based systems are not widespread.
Reply
#20
(2025-11-01, 02:29 AM)Laird Wrote: Just for clarity: LLMs are not an individual project; they're a class of AI technology, typically based on generative pre-trained transformers. There are many individual LLM projects: ChatGPT was the first one to be made public and become popular; now there are Gemini, Claude, DeepSeek, Grok, and many others.

< < < truncated > > >

I appreciate the information you are providing and the links. I might take you up on that about checking a couple of those and seeing what's up.

Now that vocabulary "generative" is in my notes somewhere on one of my units and, from memory, it was used as "generative AI" so that got on my radar screen at some point; and I think recently. Seems I remember no alarm bells going off, but maybe I was careless and missed an important clue.

But it sure seems there are a lot of players in this one and that means money is to be had and when some nutcase like me comes along and advises to slow down I am going to get hammered over the head by folks with big money in the game. In fact, it is a game that is rigged. These tech giants are pushing all sorts of stuff on folks and a lot of it isn't really needed. One of the 'jokers' in the deck is called "security" and they push stuff in the name of security and that isn't really the key point of an upgrade. It may have a very small part that could actually be fixed with an update, but they add on stuff that helps adverts be fancier and sell the security aspect and don't mention that the new thingy is 70 to 80% about that money-making adverts code included in the package.

In fact, that has probably been written up in some paper by some fearless grad student and with an edu tag search I could find something. I write "fearless" because that paper would cost that student a good job down the line at a company like Google. They see that student wrote the truth about something and the truth isn't wanted. The truth causes money flows to slow.

Maybe less than one hundred years, you think? I think I trust you on that. I might change my future warnings to reflect an update there. But I don't even have 50 years. In fact, some folks are surprised I've made it this long in this cancer battle. Funny, too, one doc back a year or so was almost sure I had only a few months. Then they tried an experimental style medicine and it worked. Maybe my "duty" is to push that warning button as much as I can. Trouble is, there are other warnings about the Net that might be needing earlier attention than this one. The Net is changing the way humans communicate, and not for the better. But that is another topic.

That is a very long *thank you* for paying some attention to this topic about the soon to be AI creatures. Might as well start calling them "creatures". Maybe. Too harsh? Maybe.
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)