TOGETHER WITH

Good morning, human brains, and welcome back to your daily munch of AI news.

Here’s what’s on the menu today:

• AI Video is getting dope 🤖 🪄

  Luma unveiled new features for Dream Machine.

• Instantly generate audio for videos 🔊 🎛️

  Google DeepMind unveiled V2A (video-to-audio).

New here? Subscribe! 😎

Peep today's ‘What Would You Do?’ at the bottom. 👇

MAIN COURSE

New AI video generation features 🤖 🪄

Yesterday, Luma AI unveiled new features for Dream Machine. It announced an Extend feature and new editing capabilities for its video generation model.

What’s up with Extend?

It allows you to increase the length of your video in style that matches your original content.

Cool. Anything else?

Luma announced upcoming video editing capabilities. You can change the background, objects, foregrounds, characters, and more in generated videos. They are “coming soon.”

Nice.

You can also remove the video watermarks if you’re in the Standard, Pro, or Premium tiers.

SPONSORED BY DROPCHAT

Spin up AI customer support agents instantly.

Dropchat is an AI-powered tool that gives your customers lightning-quick, personalized customer support (no more generic, hard-coded scripts).

With Dropchat, your chatbot learns from your company’s website, PDFs, and other internal information sources to give customers accurate answers about your specific products and services.

Companies in ecommerce, hospitality, software, and more are already using Dropchat's smart chatbots to keep their customers satisfied.

The best part?

You don't even need to know how to code to use it. Anyone can create a custom chatbot in just a few clicks.

Dropchat enables you to bring your business to life through the power of AI, creating a seamless experience for entrepreneurs all over the world.

SIDE SALAD

Automatic sound effects 🔊 🎛️

Yesterday, Google DeepMind unveiled V2A (video-to-audio). It analyzes videos and automatically generates music and sound effects for them.

What does it do?

V2A generates music, voices, sound effects, and more that “matches the characters and tone of a video.”

It doesn’t use text prompts?

You can tell it what to generate via text, but V2A is context-aware, so this step is optional.

How does it work?

First, V2A encodes the video. Then it uses a diffusion model to generate audio that matches the video/text prompts. Then it combines the audio and video data.

YOUR DAILY MUNCH

Think Pieces 🧠

Is AI just repeating words? Jonathan Marcus of Anthropic says AI models discover surprising semantic connections between concepts.

Startup News 💰

Meta paused AI training with EU and UK user data. This is due to GDPR concerns from the Irish Data Protection Commission (DPC).

NVIDIA unveiled Nemotron-4 340B. It’s a family of open models that generates synthetic data to train large language models (LLMs).

Research 👨‍🔬

OmniCorpus — a multimodal dataset of 10 billion images with text that enhances LLMs with diverse, high-quality image-text data.

HelpSteer2 — an open-source dataset designed to help align LLMs with human preferences by providing high-quality preference data.

FURRY FRIENDS

Here’s Sadie. 🐶

She’s posing for the camera. 📸

Respond to this email with your pet’s name and pic for a chance to be featured here tomorrow!

MEMES FOR DESSERT

AI ART SHOW

On Friday, I spoke with Michael Sim, Principal Founder & Head of Design of MageTCG.

He’s developing a philanthropic trading card game. The goal is to support mental health causes, military veterans, and more.

A team of independent AI artists collaborated to create the artwork. They aim to raise awareness for AI art, celebrate artists, build a community, and more.

Click here for the backstory.

WHAT WOULD YOU DO?

Hello, hero. We meet again. 😈

Choose your villain? 👇

Ideas? Comments? Complaints?

Respond to this email or hit me up on 𝕏.

Until next time 🤖😋🧠