⚡️ No, You Can't Copyright Images Made By A.I.

Welcome to our new friends and thanks for sticking around to our thousands of loyal readers!

TGIF, am I right? It's been a fast-paced week in the world of AI and machine learning, but we've managed to find some really cool stuff.

Here's what we got for you today:

• Hugging Face and Amazon Become Besties 🤗

• What is ControlNET? 🧠

• No, You Can't Copyright Images Made By A.I. 🤖

• Cool A.I. Tools and News 🤓

• A.I. Meme of the Day ⚡️

Hugging Face and Amazon Become Besties

Last month, Microsoft invested a whopping $10 billion in OpenAI, the company that created ChatGPT. Then, Google invested $400 million in OpenAI's competitor, Anthropic, which has a chat product called Claude. And now, Amazon (specifically, its cloud unit) has decided to expand its partnership with Hugging Face.

Hugging Face is a company known for its language model called BLOOM (similar to OpenAI's GPT-3). The company will use Amazon's AWS cloud infrastructure to build the next version of the BLOOM model. This means that Hugging Face will have the resources to build a competitor to ChatGPT using AWS computing power and other resources.

Last year, Hugging Face managed to raise $100 million from various investors, including basketball star Kevin Durant.

What is ControlNET?

Most people who are interested in AI have heard of tools for text-to-image creation like DALL-E, Midjourney, Stable Diffusion, and others. The quality of the generated images depends on the level of detail provided in the prompt. The more specific the prompt, the better the resulting image will be.

You've probably heard the phrase "a picture is worth a thousand words". ControlNET is a tool that works based on this idea. By using an image as part of the prompt, the text-to-image output can closely match the user's instructions.

To see how this works, check out the "Scribble" implementation. It takes a user's rough sketch, such as a lightning bolt striking a tree in my test case, and creates an image based on that plus the prompt. It's pretty cool, right? You can try it out yourself on this website and test your scribbling plus prompt skills!

ControlNET has many other image-to-image + prompt implementation methods with names like Canny, Hough, HED, Scribble, Scribble Interactive, Fake Scribble, Pose, Segmentation, Depth, and Normal map. If you want to learn more about them, you can nerd out and read the research paper here.

U.S. Copyright Office Backtracks on AI-Generated Work

Last fall, the AI-assisted comic book "Zarya of the Dawn" was granted a US copyright registration. However, just recently, the US Copyright Office reversed their decision and ruled that the artwork in the comic would not receive copyright because it was generated using AI technology. You can find the full ruling here, but here's a brief summary.

Cool A.I. Tools and News

The world of AI and machine learning is moving incredibly fast, and it's tough to keep up with all the cool new developments! Here are some of the most exciting AI-powered tools and news we've found this week.

A Meme of the Day ⚡️

Funny tweet about a fake OpenAI job posting below got me thinking. 🤔

Will AI become self-aware and conscious like humans? Some people think so, but others say "never". They say it's not just about processing power, but also involves emotions, sense of self, and other stuff that scientists don't fully understand.

What do you think? Let me know @DailyZaps

Peace out,

Daily Zap Team