Favour Y. Profile picture
Aug 29 7 tweets 3 min read Read on X
If you use AI to write, come close let me share a secret.

Most people look down on AI-written content, because it mostly denotes “low effort”.

But I’ve been writing for 7 plus years and using AI to write for 3.

Here are a few things you should do if you use AI to write.
First, always go over everything AI writes and edit them.

Most LLMs have a tendency to over explain and write in “rules of 3”.

3 items in a list (A, B, and C), or 3 sentences in a paragraph.

As a rule of thumb, you can usually remove the last sentence in most paragraphs without it affecting the message you are passing
Secondly, always do a brain dump before you write.

Don’t just say “write a tweet thread on why investment is good”

Actually pour out all your thoughts on that topic.

The more context you provide, the better AI writes for you.
Finally, you can train your AI chat bot to avoid most of the AI cliches that LLMs use.

Most LLMs have an instructions or preferences section in their settings, where you can give certain instructions on HOW you want your AI to think or respond.

Copy the prompt below and paste it in the instruction section

————————————————

Write like a human, not like an AI.

Avoid cliches and other AI sounding content.

Some words and Phrases to Avoid (AI Detection)
● Question sentences: "The result? Improved conversions." (Instead, use complete
sentences: "This led to improved conversions.")
● X/Y juxtapositions: "It's not just about features, it's about benefits." (Instead, be direct:
"Features matter less than benefits.")
● Em-dash overuse: Use periods to create separate sentences instead
● Generic business jargon: "leverage," "utilize," "synergy," "game-changer," "paradigm
shift"

● Intro phrases: "picture this," "in the realm of," "in the world of"
● False urgency: "you need to," "you must," "essential"
● Vague qualifiers: "very," "really," "quite," "actually"
● Cliché transitions: "at the end of the day," "when all is said and done"
● Command phrases: "Remember," "Keep in mind," "Don't forget"

Don't be cringe. Be Punchy, accurate, and sound like a human
—————————————-

I used this prompt to make my AI sound about 70% human for over 3 years.

If you do all this, you’ll be able to write great content that isn’t cringey and sounds like a person wrote it.

But if you want to go even deeper, to make content that is virtually indistinguishable from your voice keep reading.
The answer to this, is to teach your AI your voice.

You need to teach it HOW to sound like you.

It’s not a easy process, and it took me months to learn.

But I figured it out.

Then I started using it for my clients.

To date, I gave used it for more than 20 clients in 5 countries and written over 5 million words with AI, and none have EVER been rejected for sounding AI-like.
I put together all that knowledge over the last 2 weeks and built “AI THAT WRITES LIKE YOU

It’s a program that walks you through how to clone your voice into your AI, so that you don’t have to worry about HOW to write.

Just think of WHAT you want to say.

I launched it this week, and everyone who used it has immediately started seeing results.

If you’d like to try it out, click the link in my bio.

Also, if you’d like a 20% discount, send me a DM. Discount closes by the 31st of August
Too lazy to visit my bio?

Click here selar.com/ai-that-writes…

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Favour Y.

Favour Y. Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @FavourYusuf1

May 12
Okay, I've gotten a number of messages about how I was able to generate these images with AI, so let's talk about it.

Quick background, I've spent over 10,000 hours tinkering with AI over the last 2 years, and made thousands of dollars using AI to service clients, and I've learned a few important things.

First, is that CONTEXT is KING.

If you want to get the best results from any AI tool, especially LLMs, you need to provide as much context and details as you can.

This is very useful if you are already an expert in the field. For example, I can write a very detailed, very elaborate prompt to generate a blog post, because I have extensive experience in that field.

But what happens when you need results in a field that you are not exactly an expert in?

I'm not an expert in product photography, so how do I tell AI how to create a really good product image?

Simple. Give it enough context about what you are trying to do.

So here's the simple prompt I used:

"I want to create a luxurious product image for a crochet slippers for ladies. I have the image of the sandals, but I want to nail the copy first, then we can do the image.
So, I want you to describe the product image style and the copy, and then I will upload the image so you can edit it into that style. Again, the vibe is luxury. The image needs to communicate that luxury, alright?"

And the results were...very interesting. You can see the results below 👇🏼Image
Image
Image
Now, I have my prompt, so the next step was to simply ask the AI, in the same chat, to make the image.

I downloaded the image that was used in the original post and asked the AI to recreate it in the style guide mentioned above

And the results were...decent, but not what I wanted.

They were FLOATING IN THE AIR! 😅

I tried to correct it with this prompt:

"This is good, but the slippers seems to be floating in space, and I don't like that. Can you please remove the acrylic block and make the slippers lay flat?"

But the result was...meh

The text was cut off and it just didn't pop. (last image)Image
Image
Image
Image
So, I started in another chat, the same source image, and the same style guide.

The results were still the same. (see last 2 images below)

So, I finally went back to my prompt and found the problem.

Under "Surface and props", the prompt says "Place the pair on a low acrylic display block so they almost float above the stone."

That was the problem. The AI couldn't handle floating in space in a way that didn't defy gravity, so I simply edited that line to "Place the pair on a simple luxurious surface"

And it solved the problem.

I went back to my original chat, and made the adjustment, and instructed the AI to ensure that the text is in focus and not cropped out.

That's how I was able to get the final image I used (first image below)Image
Image
Image
Read 7 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(