Edition #28: Robots are coming for your household chores. Claude releases Artifacts.

Date: 5-Sept-2024

Hey AI enthusiast,

Welcome to the latest edition of the One More Thing in AI newsletter! Or should I call it one more thing in Robotics?

We've got some exciting updates from the world of artificial intelligence and robotics. We're covering everything from shy robot dogs in China to home helper robots that might make your chores easier.

We'll also touch on some super-fast computer chips for AI and a new model from Alibaba (Qwen) that can better understand images and videos.

There's something for everyone - let's dive in!

Best,

Renjit Philip

Founders: In case you want to stress test your AI startup idea:

In this edition:

• Deep Robotics launches shy robots

• Neo Robotics takes care of household chores

• Make your voice heard!

• New Super-Fast Computer Chips

  • The Biggest Chip Ever:

• Anthropic's New Feature: Artifacts

  • So, what are Artifacts?

• Alibaba goes open source: Qwen2-VL

• Support & Sponsor

Deep Robotics launches shy robots

Deep Robotics is a company based in Hangzhou, Zhejiang, China. I have traveled to Hangzhou a couple of times, and it is both a technology hub and a beautiful city of many gardens.

This company is creating quadruped robots that have various industrial applications. They also make robot videos (copying the approach that Boston Dynamics pioneered). Now, I find these videos funny and unsettling at the same time. We will psycho-analyze that later!

In recent news, Deep Robotics trained one of their X30 robot dogs to be 'shy' of humans and objects. The robot can recognize and detect objects approaching suddenly and autonomously avoid them. You can see how that works in practice in the video above (click the image).

Why does it matter?

Many industries face challenges like not having enough workers for specialized jobs.

Robots can take over jobs that are boring, dangerous, or hard for people to do. With this “shyness” feature, they can work safely alongside human workers.

Neo Robotics takes care of household chores

1x Technology showed off NEO Beta, their new two-legged robot made for homes. This is the kind of robot we've been dreaming about for a long time - one that can help with housework!

Some big companies are supporting this project, like OpenAI, EQT Ventures, and Samsung AI.

Why is this important?

The company wants to make lots of robots that can work with people. This might displace some jobs that don't need much training. 1x is one of the many companies trying to make robots do tedious work in homes and factories.

Maybe one day, every family will want a robot helper. The new American dream will be a robot in every home?

Make your voice heard!

I would love to know more about what you like and don’t like, specifically in connection to this newsletter. Please take this 2 min survey to help me to produce more content that is useful for you.

New Super-Fast Computer Chips

Cerebras made a new developer tool called 'Cerebras Inference'. It lets people use their special computer chips to run AI models faster.

These chips are said to be about 20 times faster than NVIDIA's chips and twice as fast as Groq's. That's impressive!

The Biggest Chip Ever:

A bit of background: Cerebras built the world's largest computer chip. It's called the Wafer Scale Engine (WSE). It's perfect for training AI models.

Faster Than Ever:

With Cerebras' technology, AI can learn much quicker. AI jobs like training, which can potentially take months, now only takes minutes.

Helping Important Work:

• Cerebras' machines are used in many areas like energy, healthcare, and big science projects. For example, their CS-2 system helps groups look at lots of information to make new medicines or study the weather.

• By letting people make bigger and more "intelligent" AI, Cerebras is helping AI better understand language and do hard-thinking jobs.

Why It Matters:

Simply put, Cerebras is another competitor to NVIDIA, for specialized use cases.

These super-fast chips are perfect for making medicines, studying life, and working with energy. They are super-smart brains that can do big AI jobs faster than regular computer parts!

By the way, Cerebras is also working with a company called G42 in Abu Dhabi to run a group of these super-smart computers.

This is definitely a company to keep an eye on! If as founders you can build AI solutions using this chip, it is an opportunity to take advantage of, right now. 

Anthropic's New Feature: Artifacts

Anthropic just released a cool new feature called Artifacts for everyone using Claude, even on phones. Now, you can make Artifacts right from your mobile device.

So, what are Artifacts?

Artifacts are like special boxes where Claude can put big chunks of information. These boxes appear in their window, separate from your main chat with Claude. They're great for:

1. Long pieces of writing (usually more than 15 lines)

2. Stuff you might want to change or use again later

3. Complicated things that make sense on their own or Information you might need to look at again

Some examples of Artifacts are:

* Documents (like school reports)

* Pieces of computer code

* Simple websites

* Special types of pictures (called SVG images)

* Charts and diagrams

Artifacts help you work with big pieces of information more easily. They keep your main chat with Claude neat and tidy by putting the big stuff in its own space.

Why does it matter?

The primary use of this feature for me at least, is to make small one-page website projects. Many people who are not technical are getting to code for the first time (I noticed this phenomenon on X/Twitter).

This is also a great way to build out MVPs as a non-technical founder. You can also "remix" other people's artifacts and build on them, which can spark new ideas and help you with a starting point.

I made a few helpful tools for people starting new businesses (there is a mix of GPTs and Claude Artifacts). You can check them out here (click button below)!

Alibaba goes open source: Qwen2-VL

Alibaba recently unveiled Qwen2-VL, a new vision-language AI model that beat GPT-4V in some benchmarks. It excels at problem-solving, math, and document analysis. The GPT-4V model they compared it to was the old model before updates, but this is still impressive.

As per their announcement: "Today we are thrilled to announce the release of Qwen2-VL! Specifically, we open-source Qwen2-VL-2B and Qwen2-VL-7B under Apache 2.0 license, and we provide the API of our strongest Qwen2-VL-72B."

Here is where you can access the models and learn more about them:

• Blog: https://qwenlm.github.io/blog/qwen2-vl/

• GitHub: https://github.com/QwenLM/Qwen2-VL

• HF: https://huggingface.co/collections/Qwen/qwen2-vl-66cee7455501d7126940800d

1. Better AI: Qwen2-VL is good at understanding pictures and words. It even beat GPT-4V (an early version) in some tests!

2. Free for everyone: Alibaba is open-sourcing two versions of Qwen2-VL for free. It reminds you of the strategy Meta adopted. If you are not leading in the LLM game, then open-source your work!

3. Useful in real life: Qwen2-VL can work with phones, IOT devices and robots (see the theme in this edition of the newsletter?). This could help make agentic software and make it easier for people to use computers and AI powered machines.

4. Understands many languages: Qwen2-VL can read words in different languages when they're in pictures. This makes it helpful for people worldwide, regardless of their language. The US based companies building LLMs are pretty focused on English language speaking population, so this could be a competitive differentiator for Alibaba.

5. More competition: Alibaba is getting into the game of LLMs with Qwen2-VL. One this is clear; there will be multiple countries with their own versions of LLMs with AI as a strategic technology.

That wraps up another power-packed edition of the Newsletter. Want the insider advantage? Subscribe to our email edition to access content before anyone else, and keep it forever in your inbox.

Also, you get exclusive content that only subscribers get access to. Discover the exclusive benefits of being a subscriber!

Support & Sponsor

This Newsletter is distributed on LinkedIn and by email. Circulation is counted as the sum of individual subscribers on both platforms.

Hit the subscribe button to receive the Newsletter as soon as it is published. If you want to interact with me on social media, follow me on LinkedIn and Twitter.

Disclosure: Some of the links in this Newsletter are affiliate links, which means we may earn a small referral fee at no extra cost to you if you purchase from them. I personally test out the AI tools during my review process. Your support helps keep this Newsletter running—thank you!