In partnership with

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

Here’s what’s on the menu today:

• Robo-butler or robo-blunder? 🛠️ 🚽

  Stardust Intelligence introduced the Astribot S1 robot.

• Humanoid Hope or Humanoid Hype? 🦾 🤖

  Mentee Robotics unveiled MenteeBot for household and warehouse tasks.

• Is this Chinese video model the Sora killer? 🎥 🍿

  China revealed Vidu at the 2024 Zhongguancun Forum in Beijing.

New here? Subscribe!

MAIN COURSE

Chores are so 2023 🛠️ 🚽

On Friday, Stardust Intelligence unveiled Astribot S1. The Chinese startup claims it demonstrates high dexterity, speed, and intelligence in household tasks.

Another bot? What’s special about this one?

The demo claims the robot is 100% AI-driven and autonomously performed actions like cooking, gardening, vacuuming, competitive cup stacking, folding clothes, and more.

Why Astribot?

The name is inspired by the Latin proverb Ad astra per aspera, which means “through the arduous journey, to reach the stardust.”

Stardust Intelligence?

It was founded in December 2022, with headquarters in Shenzhen. The founder Lai Jie previously worked with Tencent Robotics Lab, Baidu, and Hong Kong Polytechnic University.

This looks dope.

So far, the only details available are from this video and the company’s website. It’s expected to hit the market later this year.

SPONSORED BY AE STUDIO

85% of all AI Projects Fail, but AE Studio Delivers

If you have a big idea and think AI should be part of it, meet AE.

We’re a development, data science and design studio working with founders and execs on custom software solutions. We turn AI/ML ideas into realities–from chatbots to NLP and more.

Tell us about your visionary concept or work challenge and we’ll make it real. The secret to our success is treating your project as if it were our own startup.

Talk to Us Now

SIDE SALAD

Do this, bot. Do that, bot. 🦾 🤖

A couple of weeks ago, Mentee Robotics introduced MenteeBot. The Israel-based startup’s new robot is designed to tackle household/warehouse tasks.

So, what’s up with MenteeBot?

In the demo, it navigated environments, manipulated objects, and completed tasks from verbal commands. Its AI gives it bipedal movement, the ability to communicate effectively, and the capacity to learn new skills it wasn’t initially programmed for.

What’s under the hood?

It leverages transformer-based LLMs to comprehend verbal commands and plan the necessary steps to complete each task. It builds a 3D map of its environment in real-time by utilizing NeRF-based algorithms to avoid hitting things in its path.

What’s with the janky walking?

Its locomotion is powered by a Simulator to Reality (Sim2Real) approach, which translates learned movements from simulations to real-world application.

I want one. Now.

Mentee Robotics claims its production-ready version will be released in 2025.

A LITTLE SOMETHING EXTRA

The Chinese Sora? 🎥 🍿

On Saturday, Vidu was announced at the 2024 Zhongguancun Forum in Beijing. Shengshu Technology and Tsinghua University attempted to position Vidu as a competitor to OpenAI’s Sora.

What’s Vidu?

It’s a text-to-video model that generates 16-second clips at 1080p resolution. It generates videos with realistic lighting, shadows, and detailed facial expressions. It also generates imaginative, surreal content with complex scenes and dynamic perspectives.

Give me the nerdy deets.

Vidu utilizes a Universal Vision Transformer (U-ViT) architecture which allows it to simulate realistic multi-camera views of the physical world.

Is it really a competitor to Sora?

Nah. Vidu doesn’t match Sora's video output quality. For science, you can check out Sora here.

RECOMMENDED READING

The Aurorean is a clear, essential newsletter focused on demystifying science and AI. Every edition is designed to distill complex topics into understandable insights in one weekly briefing.

Discover, understand, and master the wonders of STEM with The Aurorean today. It’s free!

YOUR DAILY MUNCH

Tools 🛠️

Coconote — create notes, flashcards, and more from your recorded lectures.

Prompt library — A list of Anthropic’s optimized prompts for many tasks.

SpeedLegal — identifies risks, deciphers legal jargon, and more

mindECHO — search and retrieve data from your web content collections.

Think Pieces 🧠

Coca-Cola and Microsoft announced a partnership. Coca-Cola committed to spending over $1 billion over 5 years on generative AI, cloud tech, and more.

Drake is facing a potential lawsuit from Tupac’s estate. They’re threatening to serve Drake for using an AI-generated Tupac voice on a track.

An AI framework to generate humans in motion from prompts. It utilizes scene affordance maps to improve realism and generalization in 3D environments.

Startup News 💰

The US Government created an AI safety board. It includes the CEOs of OpenAI, Microsoft, Google, NVIDIA, and more.

Reddit launches Dynamic Product Ads. It displays relevant ads to Redditors using AI and engagement data.

Curiosity-driven red-teaming? MIT researchers developed a method to generate provocative prompts to train other AI models to avoid toxic responses.

Research 👨‍🔬

MIM4D — improves visual learning for autonomous driving by simultaneously modeling spatial and temporal details from multi-view videos.

SnapKV — optimizes KV cache usage in LLMs to process long inputs more efficiently, reducing memory usage and speeding up decoding.

MaGGIe — produces precise, consistent alpha mattes in video instance matting, optimizing multi-instance processing and enhancing guidance robustness.

MEMES FOR DESSERT

TWEET OF THE DAY

Elon’s AI company, xAI, is now valued at $18 billion.

For more info on xAI, click here.

Tag us on Twitter @BotEatBrain for a chance to be featured here tomorrow.

AI ART-SHOW

Until next time 🤖😋🧠