GrauGeist
Generalfeldmarschall zur Luftschiff Abteilung
I was going to comment on the magic serpentine belt, but the red arrow takes precedence... 
AI Fails (10 Viewers)
- Thread starter GTX
- Start date
MiTasol
Major
Jadera
Senior Airman
yay
- Thread starter
- #1,944
In Lichtenstein, they replaced the public radio by AI radio.
Warspite63
Staff Sergeant
- 1,127
- Dec 8, 2022
.. and re-yay!
BUT in a difference to the shite we usually see, this photo is a real one which has been falsified - there is an original b&w picture:
I have no idea did the AI choose the fake camouflage and insignia or did the "master" of the AI wanted to make German planes look like British?
Last edited:
GrauGeist
Generalfeldmarschall zur Luftschiff Abteilung
AI can be used as a form of image enhancement, something along the lines of a super "photoshop".
I have tried this (instead of using Photoshop tools which I don't really know how to use) and the results were similar: Insignia in wrong places, fake camouflage, windows, hatches and details that do not exist, fake guns in absurd places etc. So basically AI is like a shizo who has so many details in his brain but cannot align them and makes a mix-mash of everything.
Airborne2001
Airman 1st Class
- 184
- Jun 17, 2024
I was gonna say, because I've never seen this photo before but the AI one "felt real" even though it was AI. It turns out that it was in fact just an enhancement gone wrong, which seems to be a common case at times.
GrauGeist
Generalfeldmarschall zur Luftschiff Abteilung
I used imaging programs such as:
Paintshop Pro
Photoshop (several versions)
Turbosquid
Autodesk 3DS Max
Strata Studio Pro
All of which were used for projects like photography, web design, 3D modeling/skinning for combat sims and stuff. They required actual brain function and patients to get results.
This day and age, people want instant gratification with the least amount of effort possible and eventually they'll reap the whirlwind for doing so.
Yes, "enhancement". I think I'd call it emblemishment. Encrapulation maybe.
MiTasol
Major
Corrected
My laptop started to do ramdom crashes. I tried to work with chatgpt to diagnose the problems. But I feel it has been running around in circles for an hour. It keeps on asking for the same logging output. It did help me on one thing that I hadn't considered though, testing the graphical stack, which indicated that my problem is GPU related.
But you really have to stay on your toes to work with AI. I had to stop it from checking the same thing over and over again (caused by the limit of its "context") and don't let it draw conclusions, but draw your own. My conclusion early on in the process was already that it'sa hardware problem and I have to replace the laptop and after an hour of AI speculating, I still stand by that conclusion.
Ah well, another hour of my life I will not get back.
But you really have to stay on your toes to work with AI. I had to stop it from checking the same thing over and over again (caused by the limit of its "context") and don't let it draw conclusions, but draw your own. My conclusion early on in the process was already that it'sa hardware problem and I have to replace the laptop and after an hour of AI speculating, I still stand by that conclusion.
Ah well, another hour of my life I will not get back.
SaparotRob
Unter Gemeine Geschwader Murmeltier XIII
I like "emblemishment". Spellcheck doesn't.
Try Claude instead of ChatGPT. Claude is helping me more (and in a better way) with some PC problems. It gives better ideas sometimes and one can find some software (for testing) or some additional information easier. At least this is my experience.
ChatGPT helped me with a problem which was in fact based on a change his creators (Open AI) did: my older OS and the browsers I'm using couldn't open the saved chats. Apparently Open AI, my browser creator (Firefox) and Windows decided to make changes at the same time (end of October) which killed my chats. I asked the GPT, it told me what to do and which "secret" Windows subfolders give me the information about the failures, it wrote a letter to its own support, I sent it to Open AI and in 2 days the problem has been solved. That's team work with a human leader.
When talking to those robots, use at the end of every (or some) question commands like /human, killcritic and x10think. They can shorten the usual AI-nonsense and make the chat more helpful.
