Discover more from Magic AI Newsletter
🪄 Weekly Magic AI: Meta Connect | Spotify Podcasts | ChatGPT Updates
Top AI news of the week, Magic AI tool of the week, and article 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. This week’s Magic AI tool is an insider tip for all podcast lovers. See further below.
Top AI news of the week
🤖 Meta Connect: Chatbot Meta AI and Image Generator Emu
Meta focuses on integrating state-of-the-art AI technology into their popular apps to provide more people AI experience. On Meta Connect, they have introduced Emu (short for Expressive Media Universe), a text-to-image generation model that can create high-quality and photorealistic images. Instagram will soon have features that utilize Emu’s technology to transform photos and co-create AI-generated images. In addition, Meta experiments with forms of visible and invisible markers to help people distinguish AI-generated content.
Unlike most others in the industry, Meta doesn’t believe there will be a single super-intelligent AI that everyone uses. Meta believes in the need for different AIs for different purposes and is building AI studio, a platform for creating AIs that can assist with various tasks. Meta has also introduced a general chatbot with web access called Meta AI. This chatbot also has the image generator Emu integrated and is based on Meta’s LLM Llama 2. Besides, Meta is committed to ensuring the safety and responsibility of generative AI by working with experts.
Meta integrates AI into almost all its products, giving the world access to cutting-edge AI. The big question, however, is how Meta wants to mark AI images and texts. In addition, Meta has one advantage over the other tech companies because Meta’s LLM (Llama 2) is open-source. The open-source community is developing extensions for the Meta model (examples are 32K Giraffe or DeciLM). Meta can use these developments and findings for their own products!
Meta Connect 2023 - Meta Blog
🎧 Spotify automatically translates podcasts
Spotify is piloting a new feature called Voice Translation for podcasts, which uses AI to translate podcasts into different languages while maintaining the original speaker’s voice. This technology, developed by Spotify and leveraging OpenAI’s voice generation technology (OpenAI Whisper), aims to create a more authentic listening experience for global audiences.
The pilot includes collaborations with popular podcasters (for example Lex Fridman Podcast) to generate AI-powered voice translations in languages such as Spanish, French, and German. The goal is to deepen connections between listeners and creators and unlock the potential of human creativity through thoughtful AI approaches. This is just the beginning, as Spotify plans to expand access for more creators and languages in the future.
Language barriers are a thing of the past with this new feature from Spotify. You can hear the voices of popular podcasters worldwide, which supports the knowledge exchange. Nevertheless, the technology also offers risks in the wrong hands. Fraudsters could imitate the voices of famous people.
Spotify’s AI Voice Translation - Spotify Blog
💬 ChatGPT can hear, see, and speak
ChatGPT now has voice and image capabilities and provides a more intuitive interface for users. These features provide additional ways to use ChatGPT in various scenarios, such as having voice conversations or discussing images. Over the next two weeks, OpenAI will roll out voice and image functionality in ChatGPT for Plus and Enterprise users. The deployment of voice and image capabilities is being done gradually to ensure safety and mitigate risks associated with impersonation or fraud.
Besides, OpenAI has announced the return of ChatGPT Browsing. The chatbot now gets up-to-date access to information across the web via integration with Bing search. ChatGPT Plus and Enterprise users can enable web browsing. This feature will soon come to the free version.
This is a huge update for ChatGPT. Without a doubt! In the blog post, OpenAI mentions that its goal is to create an AGI (short for Artificial General Intelligence). In our opinion, we are still many years away from an AGI. AI works well for specific use cases and generates value there. The goal should not be to create an AGI but to add value to humanity!
ChatGPT can now see, hear, and speak - OpenAI Blog
Magic AI tool of the week
Snipd allows you to mark highlights in a podcast. This way, you can access the most valuable information from the podcast later. Currently, the tool works only for English podcasts. If you love podcasts, then try it out!
You hear something interesting.
Tap your headphones to save it to your library.
An AI generates transcripts, summaries, and titles.
Export your saved highlights to your favorite app like Notion.
Snipd Free: 2 free episodes per week
Snipd Premium: Free trial for two weeks
Article of the week
Everyone has been talking about Prompt Engineering! But it has several limitations! For example, the model always decides based on the training data. Most models use general data. They are not specific to particular use cases. If you want to create large language models (LLMs) for specific use cases, you should think about fine-tuning. In our article, we look at the LLM platform Lamini. With Lamini, you can fine-tune your LLM with 3 lines of code. Wow, that sounds incredible!
Do you want to support us as writers? You can do it through buy me a coffee. Thanks for your support.
Thanks for reading, and see you next time.
- Tinz Twins
P.S. Have a nice weekend!
Thanks for reading Magic AI Newsletter! Subscribe for free to receive new posts and support our work.
Follow us for more content: