Kind of fascinating while rather disturbing. Mostly disturbing...

 

It’s already out of control, you see a new AI impostor most every day of a personality. Regulations are needed. There is a Chinese AI moving financial markets, it’s scary what’s happening and how serious the implications really are. One day we might see an AI Trump declaring war . Made by our enemies. Now it’s easy to spot , the body doesn’t move and the mouth just isn’t right. But soon it will be 100% believable

 

Sorry ... I didn't watch the video, but I assume its content is obvious.
I do see that "Bob Crane" is written .... did they ever figure out who murd ered him?


Anyway ... yet again, we find another AI BOT invading the X204 sub-forum:
It's funny that the "new user" shows to have joined in 2007??
And only one post TODAY. ... in the "USB" post.

Many forums require a "person" to first post in a "new user" section.
Then they have to be approved (can not post anything else), and only then allowed to continue posts, if approved.
.
This is the only forum I participate in where these AI BOTs are proliferating, (because an approval is not required)

 

Whoa

I joined in 2007 and posted in the USB discussion.. Read a person's information

 

You probably missed the "AI BOT" post ... I reported it and it's now been deleted.

I have the screenshot of the post if you require to see it (and that I'm not referring to you)

 

It's sad how fast this negative side of AI is proliferating. I already think that every picture or video I see is fake until I can verify it is not. That is very quickly becoming harder and harder. Nature photographs, Jimmy Kimmel videos, you name it, they all are coming up as fake a large percentage of the time.
AI has so many wonderful and beneficial things to offer and it is being ruined by people pursuing the dark side. The whole internet started clean and was quickly polluted. The vandalizing continues.

 

Makes sense, I was puzzled


2007 was strange

 

It is indeed troubling. When I know it’s AI, 8 take it with a grain of salt. With that said, it is getting harder to differentiate AI from warm body. AI summary got me when I was reassembling after my transmission repair. I asked for torque values and was misinformed on the torque converter drain screw. I ended up over-torquing and stripped screw head. This will be a challenge if I ever need to drain it again 😒.
What if the USB post was AI trying to learn from us, the “experts”?

 

That's exactly how screws get stripped, the compilation of all the "expert" information on the internet. I am on one forum where the owner of the forum (no ads, privately and donation funded), personally deletes all incorrect information.

 

That's why they still have service manuals!

 

Retired (about five years ago) software engineer, with 40+ years experience.
I have the AI setting disabled on all my browsers (laptops, desktops, smartphones).

Me? I never trust or rely on AI-provided information, and never will.

 

Me too! Two things, truths since the early days of computers:
1) GIGO (Garbage In, Garbage Out)
2) "To err is human, but to really foul things up you need a computer." (Biologist Paul R. Ehrlich)

 

I worked on a team developing network AI for technical applications years ago before retiring. I was the earth physics guy, not the software guy. To run that stuff back then took machines at FoMCo, UNC, or DoD. The dreaming, hallucinations, and slop coming from today’s platforms has all the hallmarks of overtraining. Our networks were really good at interpolating and fair at extrapolating when trained on discrete information with high resolution. When the training included low resolution data with uncertainty, the results looked like they came from the mind of Salvador Dali. The system would never fail to give an answer. If it didn’t know a good one, it would make something up. When it had fuzzy data in the training set, the boundaries between maybe real and maybe fantasy were fuzzy too. Sound familiar?

 

Perfect analogy!

 

Nothing wrong with Dali. Eccentric is good.

 

How did *it* ==> know <== it wasn't a good one (i.e., good answer) ?

( sidenote: AI doesn't "know" anything )


So, for fun, I fired up a browser in Private mode and enabled "AI" (hahaha)

In the URL bar, I typed:

"what does AI not know"

... and tapped the Enter key.

.
Response:
AI systems are powerful tools for data processing and pattern recognition, but they *do not possess* genuine understanding, consciousness, or lived experience. While they can *simulate* human-like intelligence, they fundamentally *lack the ability* to grasp context, meaning, or nuance in the way humans do.

... thanks, I've known that for years

 

That’s probably my inability to describe it properly. The systems I worked with created a mathematical representation of how closely the result matched information in the training data base when read back in expressed as a percent. Basically: input data -> output result -> reverese process to recreate pseudo input -> compare. We used this as a kind of indicator of the confidence in the quality of the result. Overtrained networks would consistently produce lower “best fit” numbers than networks trained on high resolution data. One of my tasks was to hand filter the data in the training sets to avoid uncertainty.

note to side note: Artificial Intelligence is a misnomer. There’s no intelligence there. It’s just an algorithm even if we don’t know the structure, coefficients, and weights. My grandson wrote a philosophy thesis on can machines become truly intelligent. What I got from it was that the most basic intelligence requires the ability to to learn, the ability to create an outcome based on the learning, the ability to create an outcome based on incomplete information, the freedom to make mistakes, penalties for those mistakes ranging from inconsequential to fatal, the ability to understand that the penalty was the result of the mistake, the ability to alter the outcome from identical input, and the ability to understand that the two outcomes were related.

It’s getting late and with that last one I’m in over my head.

 

Consider the processing power of the human brain... And, it runs entirely on renewable energy!

 

I appreciate the detailed explanation... I understand much better now !!

That's an interesting thought.

Which got me thinking. Some folks in computer forums brag about their machine's (laptop, desktop, etc) "uptime". I just realized that "me" has a current uptime of over 68 years ... never tried a restart.

 

Okay. I’ll shut up. ☹️

 

I would recommend against it!

I do recommend 8 hours of maintenance time every night...

 

Why the frown face?

I appreciate the explanation ... I didn't infer anything negative.

 

Yea, makes sense ... considering I always switch the laptop to Sleep mode

( of course the laptop doesn't have human stress which ignores sleep needs "

 

Oh, sorry. I thought it was a bit of sarcasm because of the length of my explanation. I can be verbose and have been dinged by others occasionally. A lot of family members zone out around 15 words in, too. I guess I’m conditioned.

 

Understood...
I did snip a bit of the Reply, because no reason to repeat it all. As an author of five books (computer stuff), I am accused of over-explaining stuff by many folks