ChatGPT has significantly evolved since its launch, now offering capabilities that enhance—and dare I say, replace?—key daily tasks.

This issue will look at some less-known power hacks for ChatGPT, but I assume you can generally implement them on any AI chat platform of your choice.

I've tested the good, the bad, and the ugly so you don’t have to.

Here are some of my favorite hacks:

Typing “I am a business owner and this is what my company does” every time you want to get something our of GPT can be annoying and time-consuming.

ChatGPT just came out with a new feature called Custom Instructions.

Here’s how I’m using it:

Done! Now, every chat you start will use these custom instructions.

Strategically structuring your prompt for generative AI can be game-changing.

I keep learning and applying new ways every day.

In short, a better prompt = a better response

Prompt engineering is complex, and not everyone has the time or desire to master it.

This is why I partnered with Hubspot to bring you 100 free prompts, ready to copy/paste and use!

You’ll find everything from summarizing long documents to doing extensive market research, it’s all right here:

If you're like me, you have a million browser tabs open at once.

I hate how messy it gets when using many ChatGPT chats.

Until they design a better way to organize chats, I've found a solution on Reddit that works well.

This Google Chrome extension called Sider is a ChatGPT sidebar allowing you to access ChatGPT in any tab without toggling.

It supports all the major platforms like Google, Gemini, and Claude 3 (though I haven't tried those).

I’ve used it for regular text prompts, but you can chat with images.

Example: A couple of clicks on an image of a pasta dish and you can find how to cook the recipe (also haven’t tried it like that, but that’s a nifty feature!)

Everywhere you look is littered with big chunks of text written by ChatGPT and honestly, it shows.

Instead of relying on AI to write text, I use it as my diligent editor.

I usually have 'brain dumps' of long, sloppy paragraphs for this newsletter, my LinkedIn posts, emails, etc. I prompt ChatGPT to correct my grammar, and syntax, and edit for brevity.

My new favorite editing hack?

I ask it to “typeset” a large chunk of text, and it returns with neatly formatted paragraphs that are easy to scan and read.

Building on the hack above, last week I read that "delve" is the most used word in GPT-generated text.

When I read a text that clearly shows it came straight from GPT without editing, I know = written by a bad prompt.

‘Negative’ prompting tells ChatGPT what NOT to do.

And it works *chef’s kiss*

Here’s a style addition to add at the end of your prompts:

PS: When you eventually find other BS words, you can add them to your style. The above are just some of my own.

Thank you for reading! If you found this newsletter valuable, please share!

The greatest compliment I got so far was finding out new subscribers have been passing around previous newsletters around 😍