Hey brains,

Welcome back to Bot Eat Brain, the daily newsletter that smells like the Earth after a summer storm. Fresh and invigorating.

Today we're talking about two recent papers from teams at Google that will transform video editing forever.

If you like this newsletter, share it with your friends.

In this issue:

• Edit your videos using AI 🎥

• Generate entire movie scenes from text 🎬

• Instant translation that goes way beyond dubbing 💬

Interested in sponsoring a future issue? Click here to book a slot. February and March are already sold out!

Recommended Reading 📰

Your daily AI briefing

• Short form AI newsletter for busy professionals (3 min read)

• Daily AI news covering machine learning developments, AI startup funding, new AI products, and tools that let you leverage AI to grow your business.

Edit video using AI 🎥

A team from Google Research recently released Dreamix, a tool they've built to edit video just by describing what you want.

This goes beyond previous examples of text-to-video generation, which we've written about before.

By incorporating source content, this tool gives you much greater control over the final output.

And you're not just limited to video. The model is perfectly capable of creating video concepts from a single still image or even a collection of images.

The tool works by adding random noise to the input source material along with your text prompt. Both are then fed into a video diffusion model which fills in the noise based on your prompt to generate the resulting video.

Now you've got a few seconds of video, but that's a far cry from being a movie. How can you go from here to a full scene?

Generate entire movie scenes from text 🎬

Those are the prompts that were input into Phenaki, another tool being developed at Google, to generate the scene above which continues for two more minutes traveling through a variety of landscapes and situations.

Phenaki can synthesize realistic video of any length from text prompts and as you've seen it can smoothly transition between scenes and subjects as well as take camera direction.

Have an idea for a movie you just can't get out of your head?

Describe it to Phenaki and it could feasibly generate the entire thing for you.

Don't like what you get? Take out any clip and refine it using Dreamix.

Not impressed? Just wait 6 months. Video generation technology is likely to improve just as rapidly as what we've witnessed from image generation tech like stable diffusion.

TL;DR

• Google Research has released Dreamix, an AI-powered video editing tool that allows users to edit videos by simply describing what they want.

• Dreamix can generate temporally consistent videos by incorporating source content and using a video diffusion model that fills in noise based on user input.

• Phenaki, another tool being developed by Google Research, can synthesize realistic videos of any length from text prompts and can smoothly transition between scenes and subjects, taking camera direction into account.

Your next million-dollar ideas 🤑

🔮 Predictions

1. More accessible and user-friendly AI-powered video editing tools will become available to content creators, allowing them to produce high-quality video content with less time and effort.

2. As the technology continues to improve, we may see the rise of entirely AI-generated videos for various purposes, such as marketing, entertainment, and education.

3. The integration of natural language processing and computer vision technologies may lead to the development of interactive and immersive video experiences that respond to real-time user input and preferences.

💸 Business Ideas

1. AI Video Editing Services: As AI video editing technology becomes more advanced, there will likely be a growing demand for services that leverage this technology. A business could offer custom video editing services that utilize AI tools like Dreamix to provide faster, more accurate edits.

2. AI Video Advertising: With AI video editing, businesses will be able to create video ads much faster and with greater precision. A company could offer AI-generated video advertising as a service, helping businesses to create more effective and engaging ads that better target their audience.

3. AI Video Production Tools: As AI technology advances, new video production tools will likely emerge that make it easier to create high-quality videos quickly. A business could focus on developing and selling these tools to other businesses, helping them to create better content more efficiently.

Your daily munch 😋

Translate video into any language

This tool doesn't just change the audio, it changes the entire face of the actor to match that audio with 3D awareness.

Facial recognition banned in Italy

Yes. Italy banned the use of facial recognition if it is not for use against crime. This will be until new regulations or at least for one year.

Germany is introducing AI in schools

In the “AI@school” program initiated by the ministry, schools and teachers will be able to freely explore the numerous application scenarios of AI in an educational context.

Today's Featured AI-Generated Art

"The Village" - @JanFFell

Until next time ✌️

🤖😋🧠