**Top AI news of the week and Magic book tool of the week**
Hi AI Enthusiasts,
Welcome to our weekly email newsletter, where we bring you the latest updates on Artificial Intelligence (AI) in an accessible way.
## Top AI news of the week
### 🎵 Meta releases AudioCraft
[AudioCraft](https://github.com/facebookresearch/audiocraft){:target="_blank" rel="noopener"} is a kind of ChatGPT for musicians. It can generate audio and music via text. You don't need an instrument to generate music anymore. AudioCraft consists of three models: MusicGen, AudioGen and EnCoder. MusicGen, trained with Meta's own and specially licensed music, generates music from text prompts. AudioGen, trained with public sound effects, generates audio from text prompts. EnCoder should be able to generate high-quality music. Musicians and creators are the target audience. They can brainstorm and experiment with the tool.
**More information:**
* [Introducing AudioCraft: A Generative AI Tool For Audio and Music](https://about.fb.com/news/2023/08/audiocraft-generative-ai-for-music-and-audio/){:target="_blank" rel="noopener"} - Meta Blog
### 🤖 YouTube tests AI summaries of videos
Google is currently experimenting with Generative AI on YouTube to summarise entire videos. The AI summaries appear on the watch and search pages. This feature is currently only available for a limited number of English-language videos and viewers. The AI summaries are designed to help viewers decide whether or not to watch a video. However, they are not intended to replace the video descriptions written by the creators. There are already AI-powered tools for summarising YouTube videos. However, these tools have problems summarising long videos.
**More information:**
* [YouTube experiments with AI auto-generated video summaries](https://techcrunch.com/2023/08/01/youtube-experiments-with-ai-auto-generated-video-summaries/){:target="_blank" rel="noopener"} - TechCrunch
### 🦾 Google's RT-2 model
RT-2 is a new vision-language-action (VLA) model that helps robots understand and perform actions more easily. It's a model trained on text and images from the web that can output robot actions directly. The model enables robots to understand human instructions and translate them into actions. In other words: With RT-2, robots are able to learn similarly to us and transfer learned concepts to new situations. Let's look at the following example:
Push the ketchup to the blue cube (Image by Google Deepmind)
RT-2 performs well on real robot Language Table tasks. Only the blue cube was present in the training data. That's a big step for AI-based robotics research.
**More information:**
* [RT-2: New model translates vision and language into action](https://www.deepmind.com/blog/rt-2-new-model-translates-vision-and-language-into-action){:target="_blank" rel="noopener"} - Google DeepMind Blog
## Magic AI book of the week
We recommend the book "Numsense! Data Science for the Layman: No Math Added" by Annalyn Ng and Kenneth Soo to get started with Data Science.
The book provides a comprehensive introduction to the fundamentals and applications of Data Science. The authors dispense with mathematical details. It teaches the concepts, methods and techniques needed to extract valuable insights from data.
The authors use practical examples and case studies to enhance understanding. This book is especially suitable for data science newbies.
💡 Do you enjoy our content and want to read super-detailed articles about AI? If so, subscribe to our blog and get our popular data science cheat sheets for FREE.
