?? How to Create Perfect Images in 1 Minute with AI? [??? Creator's Secret Ep 1/5]

This week I am launching our new series about AI, that will unlock skills for you to be even more effective by learning how to better use it for your content creation!

PART 1: Social media weekly sum up in 1 minute flat

PART 2: Creating perfect images with AI on the first try.

?? Hi, it’s Jesse!

Every week I give you a professional tip on social media, tested and proven, no bullshit.

+ Your social media weekly sum up in just 1 minute ? (because I know you don't have time).

Join us if you want to boost your online visibility, it's once a week, we are already more than 700 members: Subscribe Now!

A new TikTok ban threat, YouTube enhances its TV app, and the imminent death of BeReal – time for your weekly sum up!

• TikTok potentially facing a ban in the United States ? Well, it's not the first time. Is it a new bluff or a genuine threat? Stay tuned... No announcement yet concerning Europe, but as I repeat every week now: don't put all your eggs in one basket, work on your presence on at least a second platform, and secure your followers' database independently!
• Launch of a new tool "Creator Search Insights " that should assist creators: the tool provides insight into the topics people are searching for on TikTok or topics that are often searched for but have few relevant videos available, worth checking out!
• TikTok is working on a new photo-sharing app similar to Instagram.
• Now it's possible to see who has viewed your Stories and to "Like" other people's posts within Stories.
• TikTok is expanding its monetization program for "AR effect creators" (TikTok’s AR) to more regions and lowering the payment threshold.

• Instagram now allows you to add up to 15 photos or videos in carousels.
• In testing: the ability to send files in the chat .

• YouTube keeps improving its TV app with a new design that makes it easier to access comments, descriptions, and purchasing functions without interrupting video playback.
• Testing a new feed for shorts called "Trending ."

• CapCut has announced that it's accepting new applications from entrepreneurs and e-commerce professionals to test CapCut Pro: Beta users will have exclusive access to test their latest e-commerce content production platform, as well as exclusive benefits and perks.

• LinkedIn is expanding its "Thought Leader " ads, now allowing brands to sponsor any organic post.
• New features and improvements added to collaborative articles .

• Pinterest is launching a new tool called "Body Type ," allowing users to select body morphologies to generate more relevant and inclusive search results (related to fashion searches).

• Once again, X has announced its intention to become a video-focused platform , the company is preparing to launch its YouTube clone app for TVs .

• BeReal, the French-made application, is on the verge of dying: The company is in a very precarious situation, facing significant challenges in attracting new users and retaining active ones, and is apparently running out of money . (RIP my baguette ????) (If you want to learn more about the situation and the story of BeReal, I recommend this well-made video by Hari )

• Meta is considering adding new perks to its paid subscription "Verified", such as the ability to add links in Reels and customer phone support. (OMG ??)

• WhatsApp could soon give us the ability to pin up to five chats at the top of the main box.

• Snapchat will allow users to save direct messages (um, isn't that the opposite of the basic concept? ??)

• Discord now allows developers to create games and applications that are accessible and playable directly within the app.

• Telegram has just reached 900 million users and is very close to becoming profitable! ??

Gen Z returning to Facebook?

Against all odds, Generation Z seems to be making a comeback on Facebook .

Surprisingly, it's mainly the Marketplace that's attracting them.

Even though Gen Z has largely stopped using Facebook to connect with friends, opting for Instagram and Snapchat instead, they are actively using its buying and selling platform, Marketplace, to save money and adopt a more sustainable lifestyle.

While Meta is striving to keep Gen Z on its platforms, with the accelerated launch of new features on Instagram for example, Facebook will likely consider promoting the Marketplace section more prominently in the future.

As for Gen Z returning to using Facebook as a social network... It's far from being guaranteed ??

?? How to Create Perfect Images in 1 Minute with AI? [??? Creator's Secret Ep 1/5]

This week, we're starting our brand new series on AI, designed to unlock your creator skills by helping you optimize your use of AI!

As promised, let's start with a first tip to help you learn how to better use AI to create your content.

I'll give you a trick to get better results for creating visuals with ChatGPT.

All without bothering you to write super long prompts.

And especially to save time, because I know you, you're like me, you don't have much time.

So read carefully until the end, because you're going to learn how to approach AI in a new way to boost yourself!

?? Note: This technique applies to the premium version of ChatGPT, but it's adaptable to other image generation AIs as well. However, if you're not already using the paid version, I highly recommend it. Considering the immense value it will bring you, $20 is nothing!

Alright, you know the drill, grab your coffee, and let's dive in!

?? Not so great visuals...

Whether you're a full-time content creator or you do it alongside your regular job, we all face the same issue: creating content takes a lot of time.

Whether you're a video creator, writer, or podcaster, creating content for social media or a brand, we all need to use illustration visuals.

But there's nothing worse than seeing "ChatGPT-like" visuals.

Because when we don't know how to write a good prompt, the result isn't very impressive.

And they tend to look alike.

Let's say there's a DALL-E style (the text-to-image AI used on ChatGPT).

If you ask it to generate an image and your prompt isn't precise and creative enough, DALL-E will tend to produce a fairly general and somewhat a messy visual.

And if you consume a lot of content, you start to develop an eye for it and quickly identify when an image has been created with AI.

Here are three examples of images generated by DALL-E with not-so-great prompts:

You can see a similarity in these visuals.

They all look a bit alike:

• Very general in their representation, the scene isn't really focused on any particular point.
• The style is quite plain, even dark, and lacks "soul."
• And they contain too many elements.

? Spending too much time writing prompts.

What we want is for it to look good and catch the eye.

But who has the time to write super precise 100-line prompts?

Because in that case: what's the point of using AI if you have to spend a lot of time on it?

Isn't AI supposed to help you? Save you time?

?? The $100K Answer

I have a question for you: who is better to understand AI than AI itself?

And now you start to see where I'm going with this.

The approach is as logical as can be.

Instead of spending hours writing or searching for the best formulas to generate your prompts, ask AI to do it for you.

Well, isn't our AI supposed to serve humanity? ??

So you might as well ask it to do that job for you, right?

?? The ultimate technique: step by step

Actually, it's super simple, and you'll probably have the same reaction as me when you start using this technique: "Of course, it's so logical, why was I making it complicated?"

(well, that's what I thought, lol ??)

Step 1: Set the context

Explain to the AI what role it should take, tell it who it is, give it life ??

Establishing the context for the AI is crucial to getting better results.

In this specific case, what you want is to write the best prompt possible.

So instruct the AI accordingly with a text like that:

You are a prompt engineer, an expert in text-to-image AI generation, with over 10 years of experience. Your goal is to write the best prompts to obtain the best possible visuals. Tell me you understand.        

This is a basic example.

You can start from this, improve it, change it according to your needs.

But it's to give you a basic understanding.

Then the AI will confirm with a somewhat message like this:

(What we want here is to be certain that the AI is taking its role into consideration)

Step 2: Tell it what you want

Next, you'll straightforwardly tell it what you want to accomplish, the task to be done.

Example:

Write the best prompt to represent a teenage mutant ninja turtle eating an ice cream in Central Park, on a summer afternoon, with an ice cream truck in the background.        

And here's the result:

You can see that you get an ultra-boosted prompt!

We've gone from my writing of not even 2 and a half lines, without much detail, to a prompt of 7 lines.

Containing many more details and ideas to include, which I wouldn't have thought of.

Or at least, it would have taken me time to think and test before being able to integrate them.

Here, in less than 1 minute, you get a nearly perfect prompt that you can use to generate an image.

And in terms of details to include in the prompt, it's clear that I wouldn't necessarily have thought to include this type of elements:

• "colorful headband and weapon belt": In my imagination, I didn't feel the need to specify this type of detail because it seems logical to me, but for an AI, it's important to indicate such details so that it can create a result that truly matches the concept of a teenage mutant ninja turtle.
• "families are strolling in the summer sun, children are laughing and playing in the green grass": Similarly, in my mind, in Central Park during summer, it's logical to have children and families, but I might not have immediately thought to mention it to the AI.
• "Make sure that the summer sunlight reflects off the buildings in the background": This is really an element I wouldn't have thought of by myself to create a summer ambiance.

But AI knows itself very well, so it knows how to 'prompt itself' to get a great result.

You just have to copy and paste this prompt.

And here's the result:

Then you can take your initial prompt and make changes if you're not entirely satisfied, for example, here I changed the element "sitting on a bench" to "standing":

Or you can simply ask it to start over from the first image produced, indicating new information, like: "start over, the ninja turtle should be standing."

You can experiment, but if the produced result doesn't suit you in terms of style, or if you notice an anomaly, it's better to go back to your initial prompt and change one or more elements.

This is how you save a ton of time ??

Because you could actually get this kind of result on your own, but how long would it take you?

You first need to generate an initial prompt.

Examine the result.

See what doesn't match or what's missing.

Then write a new prompt with the new elements.

And repeat this process multiple times before you have all the elements to integrate into your prompt writing.

Especially since there's a major issue to consider when creating images with AI: the style will never be exactly the same.

From one prompt to another, the outcome will vary greatly.

The proof is in the two images I generated earlier.

The first teenage mutant ninja turtle is in a "cute animated" style.

The second one seems more "real."

And if you have to add or remove elements in your prompts each time, it quickly becomes frustrating, because sometimes you'll like the style but not one element.

And other times it'll be the opposite.

Basically, you'll spend a ton of time and often not be satisfied.

And then you'll give up pretty quickly.

Because it's annoying not to be able to just change one element in your image and have to start over from scratch every time in random mode...

So here, you have a technique that will give you the most optimal way to get an almost perfect result every time, without having to start over 50 times.

One more time, no bullshit tips with me, my goal is to boost you!

Test it out yourself and you'll see that this trick will save you a huge amount of time and give you almost perfect results on the first or second try.

Well, it's AI, so it can mess up sometimes, but generally, you'll gain productivity ??

Alright, see you next week for the second episode of this series specialized in AI for content creators.

Subscribe if you want to continue learning techniques that will unlock your creator skills.

Until then, have a great week and take care of yourself! ??