Useful examples of LLM

seemly

Well-known member
Joined
Jul 8, 2024
Posts
192
Reaction score
347
Trophy points
64
I thought this might be a worthwhile thread to start, showcasing some useful use cases for LLM's that provide consistent results.

These could be one off tasks, tools, suggestions, anything. Even showing off how you have used LLM's recently, and explaining how you used it to achieve the end result.

Maybe it's just an article showcasing a useful 'outside the box' implementation that might otherwise have been non-obvious.

I'll start with this article, which I thought was an interesting use case.

Video scraping: extracting JSON data from a 35 second screen capture for less than 1/10th of a cent
 
Last edited:
Every Thursday, I compile and publish a 7 day "What's On Guide" for venues across our town - https://alittlebitofstone.com/stone-whats-on-guide/

I like to reduce barriers that would stop people from submitting info, so they send me private Facebook messenger messages for their venue. This is typically in text or image poster formats. I then use ChatGPT to process these messages and output the data into a consistent format of Venue, Event Name, Time(s) and Additional Info. I then copy and paste it into Excel, which has formulas in it to create the necessary information into HTML that I can then just drop into the website, and it's all correctly styled for me.

My next step will be to have ChatGPT create the HTML output as well, but as I've evolved from manually re-writing the event information out, I'm still seeing big wins in terms of accuracy and time!
 
I've used ChatGPT to create blog posts from video based interviews using YouTube's auto generated transcripts.

I wanted the transcript to read more like an article than an interview. This was to improve the accessibility of the content, as we do have at least one known deaf member at our running club.

The reason I chose for an article based write-up over maintaining the interview based structure was because ChatGPT wouldn't be able to differentiate who was talking, and when, from the transcript alone.

 
Last edited:
Back
Top