It's been awhile + Prompt hack

I've been quiet for over a month now, not because I have abandoned this newsletter.

This newsletter is where I share what I do and what I learn as and when it happens.

So for about 2 weeks I couldn't do anything due to other priorities hence nothing to share.

Then I had a pile of things I wanted to get done with before anything, so no time for writing.

I could have whipped up something with AI to send out weekly (I was so tempted at one point) but I have strictly restricted this one to be written by myself.

So that's the reason for being missing in action.

Now here's what I have been doing so far.

I started YouTube, posted a couple of videos, more coming up.
Subscribe if you are interested https://www.youtube.com/@dilini_rajapaksha

I've been working on an n8n automation for summarizing YouTube videos and save to Notion database.

It's almost complete and I'll share the blueprint soon.

I started preparing for the AWS AI practitioner certification. Will be posting more on that on LinkedIn. Reply to this email if you are interested, I will share my notes.

If you have already taken the AI practitioner exam I would love to hear from you!

Prompt Hack

If you are struggling to get consistent ideal output from LLM APIs for automations or even when just using ChatGPT casually this trick is awesome.

You can give ChatGPT an ideal result or outcome that you expect and ask it to write the prompt to get that result!

Example prompt:

I will provide you with an ideal outcome of a prompt to generate LinkedIn posts from ideas. Can you reverse engineer and give me the prompt that would produce the outcome?
Ideal outcome: [Your ideal LinkedIn post]

Now you can use the given prompt again and again!

I hope this helps,

until next time…