Writer RAG tool: build production-ready RAG apps in minutes

RAG in just a few lines of code? We’ve launched a predefined RAG tool on our developer platform, making it easy to bring your data into a Knowledge Graph and interact with it with AI. With a single API call, writer LLMs will intelligently call the RAG tool to chat with your data.

Integrated into Writer’s full-stack platform, it eliminates the need for complex vendor RAG setups, making it quick to build scalable, highly accurate AI workflows just by passing a graph ID of your data as a parameter to your RAG tool.

Learn more about our production ready RAG tooling here.

Elon Musk "Do NOT Trust Sam Altman" | SHOCKING INTERVIEW

Elon Musk discusses the rapid advancement of AI, predicting its imminent ability to create short films and personalized videos. He expresses both excitement and concern about AI's growing power, assigning a 10-20% chance of "annihilation." Musk criticizes Sam Altman and OpenAI for abandoning its open-source, nonprofit origins. 

He touches on job displacement by AI, advocating for instilling good values in AI systems. Musk also reflects on philosophical topics, including the existence of a creator and the potential for machine consciousness, while dismissing claims of encountering alien life despite widespread speculation.

WATCH THE VIDEO ON YOUTUBE

Today:

• AI Pioneers Honored with Nobel

• OpenAI Plans Independence in Infrastructure

• OpenAI Partners with Hearst Media

• Anthropic Launches Cost-Saving Batch API

• Foxconn Builds Nvidia Server Factory

• Build AI Apps with OpenAI-Gradio

• SAP Unveils AI-Powered Joule Copilot

AI Pioneers Honored with Nobel

John J. Hopfield and Geoffrey E. Hinton were awarded the 2024 Nobel Prize in Physics for their groundbreaking work on artificial neural networks. Their research revolutionized machine learning, enabling computers to process information in ways similar to the human brain. 

This breakthrough laid the foundation for modern artificial intelligence, now widely used in search engines, digital assistants, and chatbots. While celebrating their achievements, Hinton voiced concerns about the potential dangers of AI surpassing human intelligence. Both scientists’ discoveries highlight the importance of understanding and controlling these powerful technologies as they continue to shape society.

OpenAI Plans Independence in Infrastructure

OpenAI leaders have expressed concerns that Microsoft isn't providing servers quickly enough to meet their growing needs. Although Microsoft has heavily funded OpenAI’s AI development and infrastructure, OpenAI CEO Sam Altman and CFO Sarah Friar recently informed employees that OpenAI is planning to take more control over securing data centers and AI chips. 

This marks a shift in their partnership, as OpenAI seeks to reduce its dependence on Microsoft for critical resources. This change follows OpenAI's successful $6.6 billion capital raise, primarily from financial firms, aimed at supporting its future technology advancements.

OpenAI Partners with Hearst Media

OpenAI and Hearst have announced a content partnership that will integrate Hearst's extensive range of magazine and newspaper content into OpenAI's AI products. This collaboration allows OpenAI's platforms, including ChatGPT, to feature curated lifestyle and local news from Hearst’s well-known brands such as Cosmopolitan, Esquire, ELLE, and the Houston Chronicle. 

This move aims to enhance the user experience for ChatGPT's 200 million weekly users by providing high-quality, trustworthy journalism across a variety of topics. Hearst’s content will include proper citations and links, ensuring transparency and easy access to original sources.

Anthropic Launches Cost-Saving Batch API

Anthropic has launched its new Message Batches API, allowing businesses to process up to 10,000 queries at half the cost of standard API calls. This batch processing option offers a 50% discount on both input and output tokens, making AI more affordable for enterprises, especially mid-sized businesses handling large data volumes. 

While batch processing is slower than real-time, it meets many business needs where instantaneous results aren’t necessary. This shift could lead to a balance between real-time and batch AI processing, though it may also reduce focus on advancing real-time AI capabilities.

Foxconn Builds Nvidia Server Factory

Foxconn is building the world’s largest factory for Nvidia's advanced AI servers in Guadalajara, Mexico, as part of the global shift away from China. The plant will assemble GB200 Blackwell AI servers to meet growing demand, especially from big cloud companies like Google and Amazon. This move reflects efforts by Western governments to localize critical technology production. 

Additionally, Foxconn is constructing Taiwan's largest supercomputer, using Nvidia’s Blackwell chips, to advance cancer research, AI model development, and smart city technologies. The supercomputer is expected to be operational by 2025, positioning Taiwan as a leader in AI innovation.

Build AI Apps with OpenAI-Gradio

Hugging Face’s new "openai-gradio" tool allows developers to build AI-powered web apps using OpenAI models with just a few lines of Python code. The package simplifies the integration of large language models into web applications, reducing complexity and development time. 

This makes AI accessible to smaller teams and businesses without needing extensive infrastructure. Developers can customize the interface for specific tasks, such as chatbots or data analysis tools. By lowering barriers to entry, Hugging Face aims to accelerate AI adoption across industries, enabling companies to deploy powerful applications quickly and efficiently.

SAP Unveils AI-Powered Joule Copilot

SAP unveiled major AI innovations at its TechEd 2024 conference, positioning its generative AI copilot, Joule, at the center of enterprise operations. New features include collaborative AI agents capable of handling complex business tasks, such as dispute resolution and financial accounting. 

SAP also introduced the SAP Knowledge Graph, set to launch in 2025, to enhance data-driven decision-making by mapping relationships across business entities. Developer tools were expanded with generative AI capabilities and new features in SAP Build, streamlining development processes. 

🧠RESEARCH

Presto! Distilling Steps and Layers for Accelerating Music Generation

Presto! is a new method for faster, high-quality text-to-music (TTM) generation. It accelerates score-based diffusion transformers by reducing both the number of sampling steps and the cost per step. Through step and layer distillation techniques, Presto! improves efficiency, generating high-quality music up to 18 times faster than current methods.

CoTracker: It is Better to Track Together

CoTracker is a new tracking method that improves video motion prediction by tracking multiple points together, rather than individually. Using a transformer network, it accounts for the connection between points, enhancing performance. This flexible system works for long videos and can add new tracking points anytime, outperforming current methods.

UniMuMo: Unified Text, Music and Motion Generation

UniMuMo is a unified model that generates text, music, and motion outputs from any combination of these inputs. It aligns unpaired music and motion data using rhythmic patterns and employs a token-based transformer architecture. UniMuMo simplifies multimodal generation by fine-tuning pre-trained models, achieving strong performance across benchmarks in all three modalities.

VideoGuide: Improving Video Diffusion Models without Training Through a Teacher's Guide

VideoGuide is a new framework that enhances the temporal consistency of text-to-video generation models without needing additional training. It improves video diffusion models by using a guide model to refine temporal quality during inference. This method enhances consistency and image quality, offering a practical, cost-effective solution for better video generation results.

LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinations

This paper explores how large language models (LLMs) internally represent truthfulness, showing they encode more information about accuracy than previously thought. By focusing on specific tokens, error detection improves, but this method struggles to generalize across datasets. The study also reveals that LLMs may internally know the correct answer yet still produce incorrect outputs, highlighting discrepancies between internal encoding and external behavior.

🛠️TOP TOOLS

Dashworks - Create Custom AI Search Assistants

Reword - Write people-first articles with an AI trained by you.

Unriddle - Quickly find info in research papers, simplify complex topics, write with AI and keep everything organized.

Flick - AI-powered social marketing platform that help with scheduling, hashtags, caption writing, and analytics.

Theneo - Generate Stripe-like docs in seconds

📲SOCIAL MEDIA

🗞️MORE NEWS

• Meta is testing AI tools for Facebook and Instagram ads, allowing advertisers to animate still images and expand video backgrounds. These features aim to enhance ad creativity and will be widely available early next year.

• Uber is launching an AI assistant powered by OpenAI's GPT-4o to help drivers transition to electric vehicles. The tool will offer personalized guidance and incentives, launching in the U.S. next year before expanding globally.

• Adobe is launching a free web app to protect creators' work by applying attribution tags and opting out of AI model training. It integrates with Adobe's existing tools and aims for industry-wide adoption by 2025.

• Former Amazon executive Dave Clark has raised $100 million for his startup, Auger, aimed at improving supply chains for Fortune 500 companies. Drawing on his experience transforming Amazon's logistics, Clark focuses on enhancing operational efficiency.

• EdgeRunner, an Auto-regressive Auto-encoder (ArAE) for generating detailed 3D meshes with up to 4,000 faces. Using a novel mesh tokenization algorithm, the model efficiently compresses triangular meshes, improving training and generalization for point cloud and image-conditioned tasks.

• Inflection AI has introduced Inflection for Enterprise, an AI system designed for large businesses, offering full control over data and architecture. Collaborating with Intel, the system promises faster, more cost-effective AI deployment with enhanced security and performance.

• I-XRAY is a project highlighting privacy risks by showing how smart glasses, AI, and public databases can reveal personal details like home addresses or phone numbers just from a person’s face. They provide a guide for users to remove their personal information from face and people search engines like Pimeyes and FastPeopleSearch to protect themselves from potential misuse. 

• Oracle is investing over $6.5 billion in AI and cloud services in Malaysia, opening a public cloud region to boost the country’s digital economy. This initiative will support businesses in adopting AI and modernizing their technology.