AI Fails (3 Viewers)

Ad: This forum contains affiliate links to products on Amazon and eBay. More information in Terms and rules

WARSPITER said:
AI is showing the signs of degeneration which come from copies of a copy so to speak.

As the pictures posted here show, AI modelling is predicated on recursive use of data, or to put
it more correctly, going round in ever decreasing circles until it finally disappears up it's only
chance of escape.

In AI parlance this is known as model collapse. AI is artificial for sure but the intelligence part lacks
in many ways compared to humans. Looking at something such as a picture can result in the human
mind deciding something isn't right here. People can read text and say ok, i think I'll look into this
a bit more and see if there is proof to verify it.

This is why AI info and pictures start to degrade to the point of being ludicrous - long way to go yet.
Click to expand...

AI product feeding on AI input. Again, GIGO.

In one sense, our intelligence and AI is similar: both we and AI cross-check what we think is right vs what others think is right. The problem with AI is that it doesn't assess what might actually be wrong, it just accepts resource inputs into its algo and adapts it output to accommodate it, rather than saying "wait, there's something wrong here."

The fact that it cannot do this with such basics as visual information spells really bad outcomes for historiography, when we humans can cognate small data-points with some discrepancies and work to reason the differences, rather than an AI telling itself "it's all true" and simply putting out more shit that gets lapped up and repeated. You thought Martin Caiden was bad? Hold my freakin' beer.

What that really means is that AI is not actually intelligent. It is an algorithm for scanning and collecting, but not judging, weighing, and resolving.

Anyone relying on AI is trusting GIGO.
 
Thumpalumpacus said:
What that really means is that AI is not actually intelligent. It is an algorithm for scanning and collecting, but not judging, weighing, and resolving.
Click to expand...
Exactly. Any computer model that has any error automatically passes that error on to the next iteration which generally
means that more errors are introduced.
 
WARSPITER said:
Exactly. Any computer model that has any error automatically passes that error on to the next iteration which generally
means that more errors are introduced.
Click to expand...

For 60 years we took Fuchida's words at face-value about crammed decks at Midway. It took Lundstrom, Tully, and Parshall to see the problem and the latter two to research and resolve it.

That's a case study in human GIGO and how we get around it -- eventually. But does any AI actually scratch its head and ask "WTF?"

Clearly not. As a pattern-seeking heuristic, it is so far behind the evolved human brain and mind that anyone trusting it may as well ask a five-year-old about the matter.

Raising my son, I didn't give answers to him. I asked him questions that made him think it through himself where possible. In this very real sense, mistakes are the essence of learning. An algorithm that cannot spot mistakes perforce cannot learn, and that means that it is not actually intelligent. We can regurgitate mistakes too, without any digital help. :) We just try to reconcile discrepancies rather than pile them into one stool.

Anyone really, and I mean really, interested in the issue should find and read Metamagical Themas by Douglas Hofstadter, or some Nicholas Humphries. There is a lot more to thinking than these programmers seem to, errr, think.
 
Capture+_2025-11-30-21-52-30.png
 
Lucky13 said:
I've been tempted emailing YouTube about AI slop popping up there, but I'm afraid that my imaginative and colourful language would shine through in less than two sentences....and I'd get sent to my room! 😉😆😂
Click to expand...

Given the number of complaints I have seen of Spewtube "improving" uploaded videos using abysmal intelligence you would almost certainly get a reply from an AI ASSistant.
 
This sort of sh*te really has got to stop !
A number of complete tossers are presumably making money from this utter crap - not only the ludicrous images, but also the total nonsense of the narrative, being put forward as historical fact, which is altering history (for profit) to the yet to be educated, younger masses.

I've just partly watched one of those videos purporting to reveal "25 Facts you Didn't Know about ......", in this case a very well-known and respected, historically and visually accurate war movie, with a cast of top movie stars.
Screen shots from the movie, used without credit, were laterally reversed, and interspersed with totally irrelevant, ridiculous AI images, all wrapped up with a commentary droning on about total garbage, with the stated "facts" being the originator's personal fantasy - nothing stated in the video was true ..... except the title of the movie !!

Seriously, this trend needs to be classed as a criminal offence, with harsh sentences (including a f*cking good kicking !), before it gets totally out of hand.

Rant over ........ for now !
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 4 (members: 0, guests: 4)

Similar threads

CHen10
Can you use AI prompts to make aviation ww2 pictures?
Replies
3
Views
1K
33k in the air
33k in the air
O
Assistance with reaching user "Karoband" on 12 O'Clock High forum, Luftwaffe and Axis Air Forces section
Replies
3
Views
583
oreisbach
O
R
WWII Aircraft performance is down?
2
Replies
37
Views
7K
bada
B
CHen10
Help with registering at J aircraft forum
Replies
5
Views
1K
AerialTorpedoDude69
AerialTorpedoDude69
Venomstick121
Aircraft quiz
Replies
15
Views
2K
72modeler
7
Back