OpenAI's text-to-3D AI-generation tool

Plus: NVIDIA's RTNA model. Google introduces 'Magi' for search.

Hello, Engineering Leaders and AI enthusiasts,

Welcome to the 14th edition of The AI Edge newsletter. Today, we bring you Open AI’s text-to-3D generation tool. Thank you everyone who is reading this.

In today’s edition:

📐 OpenAI releases Shap·E for text-to-3D generation
🧠 NVIDIA introduces Real-Time Neural Appearance Models
🔍 Google's 'Magi' to humanize the search experience with AI
🐼 Pandas gets an AI-powered upgrade
📈 Large Language Models: Scaling laws and emergent properties

Let’s go

---

OpenAI releases Shap·E, a generative AI tool that produces 3D Outputs in seconds!

OpenAI has released a new project called Shap·E, a conditional generative model designed to generate 3D assets using implicit functions that can be rendered as textured meshes or neural radiance fields.

Shap·E was trained using a large dataset of paired 3D assets and their corresponding textual descriptions. An encoder was used to map 3D assets into the parameters of an implicit function, and a conditional diffusion model was used to learn the conditional distribution of the implicit function parameters given the input data. It has demonstrated remarkable performance in producing high-quality outputs in just seconds.

Why does this matter?

Here’s another significant contribution to the Generative AI field which can benefit a wide range of applications, including gaming, virtual reality, and augmented reality. The release of Shap·E will likely inspire further research and development in this area, paving the way for more advanced and sophisticated generative models.

Source

---

NVIDIA's Real-Time Neural Appearance Models deliver a level of realism never seen before

NVIDIA Research shared a research paper that discusses a system for real-time rendering of scenes with complex appearances previously reserved for offline use. It is achieved with a combination of algorithmic and system-level innovations.

The appearance model uses learned hierarchical textures interpreted through neural decoders that generate reflectance values and importance-sampled directions. The decoders incorporate two graphics priors to support accurate mesoscale reconstruction and efficient importance sampling, facilitating anisotropic sampling and level-of-detail rendering.

Why does this matter?

Indeed, such state-of-art AI models represent a significant step forward in the field of computer-generated, real-time graphics. It opens up the possibilities of using film-quality visuals in real-time applications such as games and live previews.

Source

---

Google's 'Magi' to humanize the search experience with AI

Google plans to move away from the traditional "10 blue links" format and incorporate more human voices into search engine results. The project, known as "Magi," will enable users to carry out natural language conversations with an AI program.

The revamped search engine will use AI algorithms to personalize search results and incorporate more visual elements to make it easier for users to find what they're looking for quickly.

Why does this matter?

The move could be an answer to Microsoft’s Bing and its AI-powered functionalities. However, since Google’s primary revenue source is ads from ‘search,’ it is more likely to act cautiously when rolling out this feature. Unlike Microsoft, they don’t have much advantage in disrupting the traditional ‘search’ experience of users.

Source

---

Pandas gets an AI-powered upgrade with Pandas AI

Pandas AI is a Python library that expands the capabilities of the popular data analysis and manipulation tool, Pandas. This library adds Generative AI features to Pandas, making it conversational and allowing users to ask questions about their data & receive answers in the form of Pandas DataFrames.

By using Pandas AI, users can easily filter & manipulate their data by asking for specific information & receiving relevant results.

Why does this matter?

With its Generative AI capabilities, Pandas AI can help:

• Developers quickly extract insights from large datasets and identify patterns to make data-driven, informed decisions.

• It can save significant time and effort compared to manually filtering and manipulating data.

• Ultimately can lead to more robust & reliable software applications

Source

---

Knowledge Nugget: Large Language Models: Scaling laws and emergent properties

Large Language Models (LLMs) have revolutionized natural language processing and demonstrated remarkable abilities in various language-related tasks. These models are characterized by several elements, including the number of parameters, the size of the training dataset, etc.

The trend of training increasingly larger models has been motivated by the discovery of scaling laws, which suggest that increasing the number of parameters is more important than increasing the size of the training set. However, recent research has shown that both elements should be increased in equal proportions.

This article provides an overview of the scaling laws for LLMs and explores the implications of these laws for training and processing data. Additionally, it discusses the emergent properties of LLMs and the challenges of accessing more complex problems using these models.

Why does this matter?

While large autoregressive language models like ChatGPT and GPT-4 have recently been successful worldwide, several challenges are involved in developing and deploying them. However, there seem to be some promising new approaches that may address some of the issues with current LLMs and potential solutions being explored by researchers and engineers in the field.

Source

---

What Else Is Happening

🍔 After beer and pizza ads, here’s AI-generated burger commercial. (Link)

🎭 Google Bard has now been released for Google Workspace accounts. (Link)

💪 Seems like Microsoft is assisting AMD to topple NVIDIA. (Link)

🧙‍♂️ Google’s ‘Magic Compose’ can craft personalized messages on your behalf. (Link)

🤖 Meta hired “highly-specialized” AI engineers from Oslo to boost its AI tech acumen. (Link)

---

Trending Tools

• Tiny nudge: Get LLM responses via email at desired frequency with a prompt.

• Netus: AI-driven tech to generate unique content 10x faster.

• Redense Sync: Build newsletters easily with a browser extension to add articles.

• Berrycast: Share screen recordings with AI-assisted summaries for project management.

• ChartGPT: Generate charts and answer data-centric questions with a personal data assistant.

• Dialog AI: WhatsApp companion with ChatGPT and voice-to-text transcriptions.

• Designed By AI: Transform design ideas into realistic images with AI.

• Xero AI: No-code analytics & ML platform with natural language data processing.

---

That's all for now!

If you are new to ‘The AI Edge’ newsletter. Subscribe to receive the ‘Ultimate AI tools and ChatGPT Prompt guide’ specifically designed for Engineering Leaders and AI enthusiasts.

Thanks for reading, and see you tomorrow.

---