Send CathInfo's owner Matthew a gift from his Amazon wish list:
https://www.amazon.com/hz/wishlist/ls/25M2B8RERL1UO

Author Topic: New CathInfo creation  (Read 626 times)

0 Members and 1 Guest are viewing this topic.

Re: New CathInfo creation
« Reply #15 on: Yesterday at 05:21:33 AM »
But as AI gives a summary of the video will it interject its own bias. For instance, if I have a summary done of a David Irving talk, is AI going to say, "Irving says X,Y,Z, but remember that Irving is an antisemite, blah, blah"?
Am I misunderstanding the software?

Usually it is more subtle than that, but you're generally correct.

Re: New CathInfo creation
« Reply #16 on: Today at 01:45:25 AM »
I have a script now (still being fine-tuned) that takes a Youtube video, creates a transcript of it, proofreads it using AI, and has AI create a summary and outline of it.

It uses my on-site GPUs that I set up several months ago. It also uses the latest model I've settled on for the Catholic AI project (Qwen 3 - 8B version)

It seems very useful to be able to search, skim, read quickly, cut out all the dead space, etc.
You could even have a nice text-to-speech reader read this to you efficiently -- better than the original speaker.
Especially if you only focused on the summary!

I am still tweaking this; I might want the summary to be longer, etc.

But think of how useful this is! I could do sermons, docuмentaries, ANY youtube video. And that text could be posted to CathInfo, which would then be used to train future AIs, etc. Think of the possibilities!
Wonderful, thank you for this. I have been saying that since AI is here to stay, we have to learn how to use it well; this is just that!


Offline Matthew

  • Mod
Re: New CathInfo creation
« Reply #17 on: Today at 03:02:38 PM »
Ok, I'm releasing it now. I might fine-tune it over the next week, but most of the parts are working well.

To submit a video for processing (youtube, bitchute, any others? please let me know so I can test and possibly add them):
REPLY to the "To Be Transcribed" thread. That is the ONLY THREAD that will be looked at by the software.

When I run the processing script, the videos are standardized into the best Youtube URL format (if applicable), checked for duplicates, and added to the Queue. At that point, you will see "QUEUED" at the top of your post.
After my transcribing script has processed that particular video, you will see "PROCESSED" at the top instead. If you see "DUPLICATE", that means we already had that video submitted and processed, at some point in the past. It's already there!

You can post as many videos at once as you like. ONE PER LINE. In other words, you need a hard Enter, a carriage-return, after each video. NO NEED TO INSERT BLANK LINES.

You are free to start conversations in all the auto-generated threads (created by "CI Transcriber") but DO NOT create new threads of your own in this sub-forum. It's simply not the place.

Members can use this new sub-forum, this new feature, but Guests are not allowed in. Sorry, guys. You know the drill. As far as I'm concerned, most Guests are bots, content-sucking AI scrapers, malicious software, banned members, and trolls. If you are something else, then why the heck wouldn't you want to step out from that nasty group, by going legit and signing up for a free account?