The secret to better AI visuals 👀

🎲 Probability of AI visuals, OpenAI accuses NYT of hacking, reviewing Microsoft Copilot's new features, and more!

Author

Everyday AI
February 29, 2024

Outsmart The Future

Today in Everyday AI
6 minute read

🎙 Daily Podcast Episode: Sometimes creating AI visuals can be a probability game. But it doesn’t have to be. Here’s our secrets to getting better results. Give it a listen.

🕵️‍♂️ Fresh Finds: AI that generates leads for your business, schools struggling with AI policies, and Big Tech investing in Figure AI. Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: Meta’s plans for Llama 3, OpenAI accuses the New York Times of hacking, and Apple shareholders reject AI transparency. For that and more, read on for Byte Sized News.

🚀 AI In 5: There's a lot of new and interesting features that have been added to Microsoft Copilot. We review them. See it here

🧠 Learn & Leveraging AI: Here’s our guide to the probability game of AI visuals and how you can win every time. Keep reading for that!

↩️ Don’t miss out: Did you miss our last newsletter? We talked about our biggest ChatGPT tips in 2024, Google CEO addresses Gemini drama, and our secret to AI-powered LinkedIn growth. Check it here!

Winning the Probability Game in AI Visuals 🎲

AI visuals are a great tool for any situation, whether you’re creative or not.

But it’s not always exactly what you wanted or needed.

If you want better AI images and videos, sometimes you have to roll the dice and play the probability game.

But you don't have to do it blindly.

Tianyu Xu, Founder of TYAI, joins us to share secrets on how you can get better AI image and video results.

Join the conversation and ask Jordan and Tianyu questions on AI visuals here.

Also on the pod today:

• Process of generating AI images and videos. 🖼
• Organizing media outputs and importance of creative rights ⚖️
• Limitations and challenges in AI video and images 😩

It’ll be worth your 38 minutes:

Listen on our site:

Click to listen

Subscribe and listen on your favorite podcast platform

Listen on:

Here’s our favorite AI finds from across the web:

New AI Tool Spotlight – AgentX generates leads for your business, BotHunt is a directory of chatbots, and Blobr is an AI assistant for business insights.

Money in AI – Figure AI, a humanoid robot startup, is now valued at $2.6 Billion after funding from tech giants like OpenAI, NVIDIA, and Jeff Bezos.

Future of Work - 39% of employees believe the current level of AI use in HR is appropriate.

AI in Education – School chiefs are struggling to create AI policies.

Read This – Here’s how the US is using AI in warfare.

1. Meta’s Plans for Llama 3 🦙

Meta's upcoming Llama 3 AI is gearing up to handle controversial queries with grace and understanding, unlike its predecessor. The new AI aims to provide context to sensitive questions, marking a significant step in AI development. Llama 3 is planned for launch in July.

2. OpenAI Says The New York Times ‘Hacked’ ChatGPT 🤯

OpenAI accuses The New York Times of using its products deceptively in a copyright infringement case that could shake up the tech industry. The tech giant claims The Times copied its works to build commercial products without permission, sparking a fierce legal showdown.

3. Apple Shareholders Reject AI Transparency Proposal 🚫

Apple shareholders shut down a request for transparency on AI ethics, raising concerns about biased decisions and privacy violations. Despite the rejection, the AFL-CIO is pushing for similar measures in other tech and media firms. Tim Cook promises groundbreaking AI developments this year, hinting at discussions on generative AI technology in the near future. 

4. US Military Uses AI Technology For Precision Strike 🇺🇸

The US military has upped its ante in the Middle East conflict by employing AI technology to conduct precise and efficient strikes on enemy targets. Utilizing computer vision algorithms and machine learning, the military can now identify and neutralize threats like underground weapons storage facilities, air defense systems, and more with pinpoint accuracy.

5. Apple Shifts Gears From Electric Car Project For AI 🚙

Apple hits the brakes on their electric car dream, bidding adieu to Project Titan after a billion-dollar investment. With Tesla losing speed and Apple eyeing user engagement gold, the tech giant shifts its gear towards AI domination.

What’s new with Microsoft Copilot!

There's a lot of new and interesting features that have been added to Microsoft Copilot.

We're giving you a brief overview of how Microsoft Copilot works and what's new with the features and UI/UX.

Or see this related video:

🦾How You Can Leverage:

Problem: AI video is so easy, anyone can do it. 

Why’s that a problem? 

Just because text-to-video and photo-to-video AI programs are easy to use, doesn’t mean the outputs will actually be … usable. 

You might end up with something like the early days of GenAI video: 

(Good luck getting those haunting AI movements outta your head.) 

But until we all get access to OpenAI’s Sora and its mind-boggling AI video capabilities, we might have to piecemeal together the best of what’s available now. 

Luckily, you don’t gotta go at it alone. 

That’s because today, Tianyu Xu showed us the ropes. 

Tianyu is the Founder TY AI, and he gave us a live walkthrough of his methodology to get the most out of AI video. It’s simple, yet so impactful that we dare you NOT to do it next you’re generating AI video. 

Ready? 

Lezzzz get it 🎬

1 – Anyone can do it 🧑‍🎨️

Can you type? 
If so, you can create AI videos. 

Tianyu creates amazingly detailed AI videos, yet his background wasn’t even in the creative field.  

With a market research background, Tianyu had a passion for understanding and improving methodologies.

So when he first started using AI image and video creation tools, he found he had a knack for breaking down the creation process. (You know, like a market researcher.) 

Try this:
If you’re unsure you can create something from scratch, let us guide you. 

We actually created a 3-hour course you can do at your own pace.

The course is aimed at beginners and walks you through, step-by-step, on creating a marketing campaign using ChatGPT, Bing Chat (now Copilot), Midjourney, Runway and Eleven Labs. 

It’s legit a must for anyone interested in learning or improving GenAI skills. 

Just reply to this email with “Course” and we’ll send ya all the info.

2 – Test intentionally 🤔

Smashing the regenerate button 50 times won’t help you in the long run. 

(Although it might get you a really cool picture of a cat kayaking a waterfall, like we talked about on today’s show.) 

Tianyu broke down his process of probabilities and rolling the dice.

In short, once you understand how AI products like Midjourney, Dall-E, Runway and Pika labs work, then you can better understand what to test when ‘rolling the dice.’ 

Try this:
Copy Tianyu’s methodology — he’s spent countless hours experimenting and the process can save you hours.

To get started, he brought in an AI-generated image into Pika Labs, and used Pika’s image-to-video generation feature. 

Here’s the tl;dr version of rolling the dice. 

  • Know that each software (Midjourney, Dall-E, Runway, Pika Labs) handles photo and videos differently

  • Start testing high level and run many generations (a generation is just a variation of an output) 

  • As an example, on today’s show, Tianyu first tested (live) directions of camera movements inside of Pika Labs. 

  • He did one generation per direction for simple camera movement (left, right, up, down)

  • Then, he tested the motion speed, trying out a 1 (little motion) all the way through a 4 (a ton of motion) 

3 – Double down on what works ↘️

Roll the dice. 

See what you like. 

Double down. 

It sounds simple enough, but we’ve seen and read of AI aficionados going about the process haphazardly.

In the end, Pika Labs and other AI generators are just that — generators. Results may be random and a bit hard to control. 

That’s why Tianyu said you have to start with simple, foundational movements and intentionally test how the system may treat movement, or interpret how motion should occur in each generation.

It won’t always be pretty. But, if you take it step by step and double down on what works, you can achieve some great outputs. 

Try this:
Wanna know more about AI video tools and how they may impact the future of creativity? 

We did a deep dive here, so check that episode out. 

Also, with the impending impact of OpenAI’s Sora, you’ve gotta keep an eye out on the impact that may have.

Tianyu went into some IMPRESSIVE depth breaking down Sora prompts, so go check that out. 

⌚

Numbers to watch

$2.6 Billion

Figure AI, a humanoid robot startup, is now valued at $2.6 Billion after funding from tech giants like OpenAI, NVIDIA, and Jeff Bezos.

Now This …

We wanna hear from you!

Vote to see live results

What AI visual software do you wanna learn more about on the show?

Login or Subscribe to participate in polls.

Reply

or to participate.