🐙 Top ways to use ChatGPT for repurposing your podcast (and grow your audience)
I’m building an email series that will cover practical ways to help podcasts grow and get repurposed more easily. If you’d like full info once the item is ready, fill out this brief interest form.
hiya Coffee Cup,
The older I get, the more I appreciate the simple things.
Like oven roasted vegetables, squishy slippers and tight hugs.
And knowing more about the tools I’m using. Like ChatGPT 3.5.
I’ve read a bit about what info the tool was trained on (which means the info it was fed to be able to respond to us) but I wanted to ask more. Find out more. Now that I’ve been using it for awhile I want to know why it’s quirky in the way I don’t want it to be.
And get some tips from Chatty itself to work around those quirks.
So that’s exactly what I did in the episode today.
Full details below.
But first, I want to share someone else with you.
Prompting guide
Yes, yes, this prompt guide is for Gemini, but honestly, it maps over to ChatGPT 3.5 and other LLMs just fine. The four main areas this post (and the guide it’s talking about) over are: persona, task, context and format. PTCF doesn’t flow, does it? Hmmmm. Need to think of an acronym for this because this is a super clear way of creating a clear prompt.
PaTCliF? whatever that is.
CuTPuF? ugh, these letters don’t make this easy. one more try….
FliPCuT? too many extra letters.
Can you think of one?
And special thanks to my husband, who sent me this post.
What info is ChatGPT 3.5 trained on?
As promised, here is the episode about what is IN ChatGPT 3.5, according to the voice prompt conversation I had with it today.
Listen to the episode:
audio and transcript (above) AND on all podcast apps under “Coffeelike Experiments…”
YouTube: click on the video below
Other Episode info
👉🏾This episode is sponsored by Streamyard the very platform that we record, go live and distribute these videos to YouTube, LinkedIn, Twitter, and Facebook.
🤸🏽 Music from Key Frame Audio (in podcast version sometimes): https://keyframeaudio.com
👀 Main takeaways below!
Proud to be part of these events in 2024!
April 19, 22-24th Podfest Masterclass: AI & Creator Tools Summit
(my session is on April 23rd)
🤗Full schedule: https://podfestexpo.com/master-class/schedule/
✍🏻FREE ticket: (use code "Fuccio50" for your discount)
and
June 19th: The International Women’s Podcast Awards.
I’m a judge this year. Not sure if I’m going to the in person event on this date yet. If you’re going, let me know.
😍 If you enjoyed this post, hit the like button above.
Episode Main Takeaways:
ChatGPT 3.5, according to this conversation, is trained on (aka fed) text. written texts. not audio (yet?)
Chatty has access to a lot of publicly available social media posts (before January 2022)
there’s a way you can get less formal language out of Chatty
you can ask it to reply in shorter “speaking” turns
(you’ll have to listen to how to do these two tips)
sometimes Chatty just stops working
or
there is some secrecy about the diversity of the texts it was trained on
Which do YOU think it is?
Other Coffeelike Media Productions
Solo Worklife Meditations: relax and mentally shed your work worries
Geopats: conversations about what grounds us in a different culture
Global Podcast Editors: podcast editor challenges from around the world
If you have any ChatGPT prompt or task you want me to try out, comment below.
Cup out,
Steph
13: What info is IN ChatGPT 3.5 exactly?