Well, I still dislike so-called "AI" because it WILL confidently "make stuff up", and with complex topics it hard to differentiate between useful directions vs misleading directions.

I consider AI to be a good tool for routine tasks. As for areas where the AI user is not a specialist, the results should be treated with extreme caution and verified by other means. However, the development of AI tools is currently very rapid; free models are far superior to those from just a year ago, and paid models noticeably surpass free ones. But they should still be used with the same caution. For example, AI currently handles Bash scripts and configuration files for various applications very poorly, while making fewer mistakes in Python and Rust apps.

Best regards,
Andrey


пн, 16 мар. 2026 г. в 20:07, Andrew Randrianasulu via Cin <cin@lists.cinelerra-gg.org>:


пн, 16 мар. 2026 г., 19:35 Phyllis Smith <phylsmith2017@gmail.com>:
Terje, I checked out a few of your Google-ai urls. 

Well, I still dislike so-called "AI" because it WILL confidently "make stuff up", and with complex topics it hard to differentiate between useful directions vs misleading directions.

https://en.wikipedia.org/wiki/Transformer_(deep_learning)

*technically* it surely interesting history, but socially ... this is Trainwreck, and those trillion-dollar companies will happily trade your health (at very minimum) for their own financial gain 

It also uses resources (from water to electricity to memory chips) but solving this on consumer side is nearly impossible because "99%"  just eat whatever fake innovation Big Tech pushed on them.  I suspect this lack of self-control/caution does MUCH more damage than "just" consumerism, but because there is no "will pill" I'll stop here. 



Quite interesting. Thanks for passing them along. (Also thanks for doing the "snips" because I do not always get the longer emails released from being held when over the limit in a timely manner and, after all the snipped parts are already in the thread.)

Beside for fun, I have played a bit with Google-ai to see what answers are generated to questions in the same genre - for what they are worth or not. Hopefully my shortened urls to save space here will work:

CINELERRA-GG GIT rpm.bld
https://tinyurl.com/2ebt43us

cinelerra.spec vs build script
https://tinyurl.com/35k6493w

what is included and not in Suse Tumbleweed - cinelerra-5.1-_.susetw.x86_64.rpm
https://shorturl.at/EraKS


how to configure cinelerra-gg with unbundled ffmpeg
https://tinyurl.com/4sawurxe

cinelerra configure --enable-ffmpeg
https://shorturl.at/Moi7F

what is cinelerra-gg thirdparty
https://tinyurl.com/3p2xbpx7


see what cinelerra-gg appimage is configured with
https://tinyurl.com/3m8xtrvm

is cinelerra-gg rpm static or dynamic linked
https://tinyurl.com/4w7nxw7j

how to find dynamic linked libs on linux
https://tinyurl.com/ywwkfuyn


should make clean be used after make errors
https://tinyurl.com/2z9j4nzf

_______________________________________________
Cin mailing list -- cin@lists.cinelerra-gg.org
To unsubscribe send an email to cin-leave@lists.cinelerra-gg.org