Any-to-Any Generative AI Model, Generative AI for 3D game development, Mobile app Development from Text, Context Window of 1 Billion+ tokens, Bark on Discord and more

Greetings and welcome to this week's AI Brews - your thoughtfully curated guide to AI products, learning resources and a concise roundup of the week's impactful news. Our goal? To provide a balanced selection in the rapidly evolving AI landscape, keeping you well-informed without the information overload. We value your feedback - don't hesitate to reply to this email with suggestions on how we can make this better for you. Thanks!

In today’s issue:

1. AI Pulse: News, Insights and Social Spotlight of the Week

2. AI Toolbox: Product Picks of the Week

3. AI Skillset: Learn & Build

---

🗞️🗞️ AI Pulse: News, Insights and Social Spotlight of the Week

🔥 News & Insights

1. Microsoft Research present Composable Diffusion (CoDi), a novel generative model capable of generating any combination of output modalities, such as language, image, video, or audio, from any combination of input modalities. Unlike existing generative AI systems, CoDi can generate multiple modalities in parallel and its input is not limited to a subset of modalities like text or image.[Details].

   Loading video

2. MoonlanderAI announced the alpha release of its generative AI platform for building immersive 3D games using text descriptions [Details].

3. Bark, text-to-audio model, is now live on Discord. Bark can generate highly realistic, multilingual speech as well as other audio - including music, background noise and laughing, sighing and crying sounds. [Details | GitHub].

4. OpenAI's Code Interpreter plugin, allowing ChatGPT to execute code and access uploaded files, will roll out to all ChatGPT Plus users within a week. It enables data analysis, chart creation, file editing, math calculations, and more [Twitter Link].

5. OpenAI announces general availability of GPT-4 API. Current API developers who have made successful payments can use it now, and new developers will have access by month's end [Details].

6. Microsoft AI presents LONGNET a Transformer variant that can scale the sequence length to 1 billion+ tokens without sacrificing performance on shorter sequences [Details].

7. Researchers present a neural machine translation model to translate the ancient language Akkadian on 5,000-year-old cuneiform tablets instantly to english [Details | Paper].

8. A set of open-source LLM models, OpenLLMs, fine-tuned on only ~6K GPT-4 conversations, have achieved remarkable performance. Of these, OpenChat-13B, built upon LLAMA-13B, is at rank #1 of open-source models on AlpacaEval Leaderboard [GitHub |Huggingface| AlpacaEval].

9. Researchers have developed an AI tool named CognoSpeak that uses a virtual character for patient interaction and speech analysis to identify early indicators of dementia and Alzheimer's disease [Link].

10. Secretive hardware startup Humane, shares details about its first product: ‘Ai Pin’. It is a wearable, AI-powered device that performs smartphone-like tasks, including summarizing emails, translating languages, and making calls. It also recognizes objects using a camera and computer vision, and it can project an interactive interface onto nearby surfaces, like the palm of a hand or the surface of a table [Details].

11. Nvidia acquired OmniML, an AI startup whose software helped shrink machine-learning models so they could run on devices rather than in the cloud [Details].

12. Cal Fire, the firefighting agency in California is using AI to fight wildfires [Details].

13. Over 150 executives from top European companies have signed an open letter urging the EU to rethink its plans to regulate AI [Details].

14. Google updated its privacy policy: the company reserves the right to use just about everything users post online for developing its AI models and tools [Details].

15. OpenAI believes superintelligence could arrive this decade. Announced a new project, Superalignment with a focus on aligning superintelligent AI systems with human intent [Details].

🔦 Open Source Projects

1. Embedchain: a framework to easily create LLM powered bots over any dataset [Link].

2. GPT-author: uses a chain of GPT-4 and Stable Diffusion API calls to generate an an entire novel, outputting an EPUB file [Link],

3. GPT-Migrate: Easily migrate your codebase from one framework or language to another [Link].

---

🔍 🛠️ AI Toolbox: Product Picks of the Week

💎 FlutterFlow AI Gen: Generative AI features are now in FlutterFlow, the no-code app builder. From generating database schema and pages to creating components, color themes and custom code - all from text descriptions.

💎 Wondercraft: Turns written content into Podcasts with hyper-realistic AI voices or your own cloned voice.

💎 Ween.ai: An AI-powered platform for user research that turns customer qualitative data (interviews, feedback) into actionable insights.

💎 Twine Ambient: Uses AI to automatically summarize Zoom recordings, Slack channels, news articles etc., and distributes these updates via a single feed.

---

📕 📚 AI Skillset: Learn & Build

• Building AI Products with OpenAI - a free course by CoRise in collaboration with OpenAI [Link].

• Gartner experts answer the top Generative AI questions [Link].

• Deploy LLMs with Hugging Face Inference Endpoints [Link].

---

AI Brews is free, and your sharing it with a friend helps us grow. Thanks for your support and have a nice weekend! 🎉 Mariam.