I don’t think there is an administrative lock on editing. Feel free to put that part back if you wish. Like I said, as long as you stand behind what is posted, no problem. We just want things to be factual with as little guess work as possible.
1 Like
Text edits here are quite rare and usually limited in scope. It looks like Eelik did not recognize the boundaries of the AI stuff. But the forum software automatically logs all edits (probably puts it in GIT or similar). It was easy to copy some text from the first revision and put that back again.
Yep, sorry about that. I’ll look into the editing history. I still stand behind removing AI stuff. There’s more to it than just “possibly helpful info”. At the moment the burden of proof lies on them who add any AI made stuff anywhere for any purpose without first knowing how it affects the whole humankind and our lives. I may be overdramatic, but the topic has been discussed everywhere from many angles.
We want original textual content on this forum. This has been our consensus.
EDIT: the poster seems to have found the history and restored his original contribution without the AI stuff, which is good. My apologies again.
3 Likes
Over on the Virtualmin forum it is out right banned. The exception being language translation.
It’s bad enough when humans post hallucinations. There is one popular website that people find for advice on Webmin/Virtualmin that simply has wrong info and the W/V team have asked they modify the information and they simply refuse. (SEO whores?) Those posts get deleted too so as not to pollute the forum with incorrect information because search engines find it and people don’t read entire threads. ‘It’s on the official site so it must be correct, right’?
So, yes, it isn’t a bad idea to remove ANY incorrect/harmful information.
1 Like
It’s offtopic (but this thread has gone offtopic anyway, and I really don’t mind) - but I think this probably depends on what AI is being used, and frankly, I don’t think it’s accurate for most AI used today. The main reason is that AI in the sense of LLM’s really doesn’t work that way. It doesn’t start with something existing and then modifies it if you ask it to generate some code. It basically starts with random noise and then starts refining that towards an output that’s deemed a good match with existing training data. The net result can indeed be code that strongly resembles something already existing, especially for common/routine tasks on common platforms. It’s sometimes difficult or even impossible to distinguish parallel invention from plagiarism. This issue predates Ai/LLM’s.
1 Like
Guys is time to learn how to discern AI writing style and don’t engage with trolls.
@koraks In my case that deleting completely previous installations and not using antialising (no antialiasing setting) on older hardware makes the whole experience better.
In Windows, If you use its antivirus, it ‘‘likes” to scan kicad libraries and it takes quite a bit of resources.
In such case… I add exceptions so it stop hogging resources all the time when working with kicad.
YMMV
And that is the entire problem. The training data isn’t vetted. It’s sucking garbage off the internet. Even before 2021 which is the last year Internet data is considered safe from AI, there’s tons of defective content. Now the AIs are desperate to find more training data and they absolutely slurp up AI wrong think into more AI wrong think.
The result is garbage in, garbage out.
6 Likes
Thanks; I already had experimented with the antialiasing settings and they didn’t make a difference for me.
However, I’ve had another look at the display drivers and found that HP offered a driver that was somewhat newer than the one I had been using previously. As I wrote above, I had been using the latest drivers that Intel offered, but apparently HP have released some drivers after Intel stopped supporting this chipset. That newer HP-distributed driver did the trick; everything is now running as smoothly as can be. The driver version I’m currently running is 20.19.15.5171; previously I was running version .5126 or thereabouts.
7 Likes
Late to this party….
I have a 13 month old new PC and a 1 month old new PC. Windows 11. They both go “out to lunch” for a few seconds a couple of times a day. Both machines fully updated. Both are >1 monitor. Graphics cards are Nvidia RTX A400. Processors are either 12 core Xenon or 16 core i9. Memory is either 64GB or 128 GB.
This happens no matter what I’m running. Occam’s Razor has me thinking that if it’s between a KiCad problem or a Windows 11 problem, I’d go with a Windows problem.
I wonder if it’s the auto-backup mechanism for projects
Are your projects on a network share?
1 Like
Recent Win11 versions tries very hard to push home user files onto the cloud, which will be worse than a network share for latency
I’ve looked at auto-backup, not sure that’s it though. I don’t have anything on the cloud, the “One Drive” (or whatever microsoft calls it) is off, and when it locks up everything locks up, even cursor movement. I would hope that being in an editor (eclipse) and typing would not cause the broswer (chrome) to lock up and lock up the cursor too.
It smells like a garbage dump on some system cached table. You would think a modern OS wouldn’t have those problems, FreeBsd Unix is good for millisecond real-time response. But then again, it’s windows. Neither of these machines are “Hello Kitty” i3 class. a 3 to 4 second lock up is enough time ot do a lot of things.
At any rate, wanted to pass this on. It’s not easy to figure this out, and I see lag on many programs not just KiCad.