Diversity In Stock Footage

Hello Friends,

We’re here with another edition of the Pentridge Newsletter. Let’s get up to speed with the latest in AI. This is what we’ll be getting into:

• AI Training and Recent Advancements

• Diversity in B-Roll and Stock Footage

• Quick Fire AI News

Utilizing GPTs in ChatGPT Workflows

In the last week OpenAI added a new feature to ChatGPT that it’s beginning to roll out, to where now you can call for GPTs in the middle of your prompt workflows. 

This is how it works:

1. Prompt ChatGPT with a question.

2. Add a GPT to the conversation by typing "@" and selecting one. 

3. Now you can more easily streamline your tasks without having to manually search for the GPT

It’s little tweaks like this that can go a long way to very quickly increase your productivity. 

Google's Lumiere for Realistic AI Videos

Google has revealed Lumiere, an AI that generates 5-second video clips with unprecedented realism and consistency. It works by utilizing an advanced "space-time" neural network to create the entire video in one pass. This allows for greater coherence than other models that generate distant keyframes.

Lumiere can create videos from text prompts or animate still images. The advances suggest continued rapid progress in AI-generated video during 2024 and competes with existing top of the line AI video tools like RunwayML and Pika.

Creating Culturally Relevant B-roll with AI Tools

When it comes to content creation, a lot of us would agree that diversity and cultural relevance in visual media are important yet often challenging to achieve with conventional stock resources. AI tools like Midjourney offer an effective solution by enabling creators to generate custom B-roll tailored to their specific needs.

The Process:

1. Starting Point: Begin with an existing image that somewhat aligns with the desired visual narrative. This image serves as a foundation for what's to come.   

2. Transformation: Using midjourney, you can create an image inspired by the original , morphing it into a version that better reflects the desired diversity and context. For instance, an image of a man sitting down drinking a cup of coffee can be changed to a woman doing the very same thing with a quick prompt.

3. Execution: The transformation is executed within Midjourney's user-friendly interface, currently accessible via midjourney’s Discord channel. Users input a descriptive prompt, influenced by the original image, to guide the AI in generating the new, culturally nuanced visuals.

4. Result: The outcome is a selection of images that not only meet the diversity criteria but also maintain the thematic essence of the original. These AI-generated visuals can then be seamlessly integrated into video content, enriching it with culturally relevant B-roll.

This method not only democratizes access to diverse visual content but also empowers creators to challenge biases and stereotypes in media representation. By leveraging AI, content creators can ensure their work resonates more authentically with a broader audience.

Check out a full video breakdown of creating diverse stock footage:

Quick Fire AI News

1. USPS Modernizes with Generative AI

2. OpenAI & Arizona State University Partnership

3. Microsoft's AI Innovations

4. Major AI Research Partnership

5. Google Aims for World's Most Advanced AI by 2024

6. The AI Arms Race Heats Up

7. Altman Seeks Billions for AI Chip Factories

8. Zuckerberg Plans Open Source AGI

9. Figure Brings Robots to BMW Factories

10. FDA Clears AI for Skin Cancer Detection

11. Samsung Launches AI-Powered Galaxy S24

12. AI Takes Center Stage at Davos 2024

13. Apple to Run AI Models on iPhones

 ^{Today's email was brought to you by Aki}

^{Was this email forwarded to you? You can subscribe here!}