Meta's New AI model, Generating websites and 3D textures using natural language, economic potential of Generative AI 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. ElevenLabs has launched AI Speech Classifier - an authentication tool that lets you upload any audio sample to identify if it contains ElevenLabs AI-generated audio [Details].

2. Nvidia Research presents SceneScape - a method to generate long-term walkthroughs in imaginary scenes just from an input text prompt [Details |Paper ].

   Loading video

3. Meta AI introduces the Image Joint Embedding Predictive Architecture (I-JEPA), a new AI model which learns from the world like humans and excels in computer vision tasks, while being more computationally efficient. It learns by creating an internal model of the outside world, which compares abstract representations of images (rather than comparing the pixels themselves). It can also be used for many different applications without needing extensive fine tuning. Meta is open-sourcing the code and model checkpoints [Details |Paper].

4. Meta wants to make the next version of LLaMA, its open source LLM, available for commercial use [Details].

5. Adobe launched Generative Recolor, a new tool powered by Adobe Firefly generative AI that lets you generate custom color schemes using texts prompt like “strawberry fields,” “faded emerald,” etc. [Details].

6. OpenAI announced:

   1. new function calling capability in the Chat Completions API

   2. updated and more steerable versions of gpt-4 and gpt-3.5-turbo

   3. new 16k context version of gpt-3.5-turbo (vs the standard 4k version). 16k context means the model can now support ~20 pages of text in a single request.

   4. cost reductions: 75% on embeddings model and 25% cost on input tokens for gpt-3.5-turbo [Details].

7. Meta AI released MusicGen - an open-source music generation model that can be prompted by both text and melody. See here for generated samples and comparison with Google’s MusicLM and others [Paper | Huggingface Demo | GitHub].

8. McKinsey published a report ‘The economic potential of generative AI: The next productivity frontier’ . The report estimates that generative AI could add the equivalent of $2.6 trillion to $4.4 trillion annually across the 63 use cases. About 75 percent of the value that generative AI use cases could deliver falls across four areas: Customer operations, marketing and sales, software engineering, and R&D [Details].

9. EU lawmakers pass AI regulation, requiring generative AI systems, such as ChatGPT, to be reviewed before commercial release. It also seeks to ban real-time facial recognition [ Details].

10. Google Lens can now identify skin conditions. Lens will also be integrated with Bard, Google’s AI-powered chatbot, enabling Bard to understand images in user prompts [Details].

11. AMD announced its most-advanced GPU for artificial intelligence, the MI300X, which will start shipping to some customers later this year [Details].

12. Vercel introduced Vercel AI SDK - an open-source library to build conversational, streaming and chat user interfaces. Includes first-class support for OpenAI, LangChain, and Hugging Face Inference [Details].

13. Vercel announced 'Vercel AI Accelerator, a 6-week long accelerator program with $850k in free credits from OpenAI, Replicate and others [Details].

14. Salesforce announces AI Cloud - generative AI for the enterprise. AI Cloud includes the new Einstein Trust Layer, to help prevent large-language models (LLMs) from retaining sensitive customer data [Details].

15. Cohere and Oracle are working together to make it easy for enterprise customers to train their own specialized large language models while protecting the privacy of their training data [Details].

16. Coda released Coda AI - the AI-powered work assistant integrated in Coda to automate workflows. Coda also announced ‘Coda's AI at Work Challenge’, offering $40,000 in total prizes to the makers who submit the most useful Coda AI template to the Coda Gallery [Details].

17. OpenAI, Google DeepMind and Anthropic have committed to provide “early or priority access” to their AI models to UK in order to support research into evaluation and safety [Details].

🔦 Social Spotlight

1. How people using LLM-written code auto-add malware themselves [Link].

2. An ER doctor shares how he’s using ChatGPT to help treat patients [Link].

3. Announcing Prem — Private Open Source LLMs for ALL [Link].

4. How to generate Artistic QR codes [Link].

---

🔍 🛠️ AI Toolbox: Product Picks of the Week

💎 Poly: A web-based tool for generating high-resolution 3D textures using a simple prompt (text and/or image). No sign-up is needed to try the free plan.

💎 Framer AI: A new generative AI tool in Framer (no-code website builder) for designing and generating websites using just natural language prompts.

💎 Receiptor.ai: Connects to your email inbox to automatically identify and extract data from receipts. Converts the extracted data to PDF/CSV formats.

💎 Composer: An AI-powered platform to create, test, and execute investment strategies without needing any technical skills.

💎 Backengine: Lets users create, test and instantly deploy backend APIs powered by large language models without any code or infrastructure.

---

📕 📚 AI Skillset: Learn & Build

1. A Practical Guide to Midjourney Commands [Link].

2. Introduction to Vertex AI Model Garden [YouTube Link]

3. How to use the Chat Completions API in combination with external functions to extend the capabilities of GPT models [Link}.

---

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