The most recent update, covering developments from July 17th, 2023, showcases significant progress and announcements in the fields of data science, machine learning, and artificial intelligence. This week's focus centers on Anthropic, Stability AI, Microsoft, Meta and xAI.
Anthropic Launches ChatGPT Rival Claude 2
Anthropic has introduced Claude 2, a new model with enhanced capabilities in data science, machine learning, and artificial intelligence. Claude 2 offers improved performance, longer responses, and can be accessed via API or the beta website, claude.ai. Users praise Claude for its ease of conversation, clear explanations, reduced potential for harmful outputs, and a longer memory. The latest model showcases advancements in coding, math, and reasoning, achieving a remarkable 76.5% score in the multiple-choice section of the Bar exam, surpassing Claude 1.3's 73.0%.
Additionally, Claude 2 scores above the 90th percentile on GRE reading and writing exams, and similarly to the median applicant on quantitative reasoning. Claude 2 is presented as a friendly, enthusiastic colleague or personal assistant capable of natural language instructions for various tasks.
The API for businesses is offered at the same price as Claude 1.3, and the beta chat experience is available to users in the US and UK. The model's input and output length have been increased to accommodate up to 100K tokens, enabling it to handle extensive technical documentation or even full books.
Stability AI Releases Stable Doodle, a Sketch-to-Image Tool
Stability AI, the company behind the image-generating model Stable Diffusion, is introducing Stable Doodle, a new service that transforms sketches into visually appealing images. Using the latest Stable Diffusion XL model, combined with Tencent's T2I-Adapter control solution, Stable Doodle allows for more precise control over image generation.
This tool is aimed at both professionals and beginners, enabling anyone with basic drawing skills to create high-quality original images in seconds. Users can provide a sketch and a prompt to guide the image generation process, with Stable Doodle offering 14 styles of art.
While the service is free but limited, Stability AI envisions it as a valuable tool for designers, illustrators, and other professionals to save time and maximize efficiency in their work.
Meta AI Reveals CM3leon, an Advanced Text-to-Image Generative Model
Meta AI has introduced CM3leon, a novel multimodal model combining text and image production. This model is the first of its type, using a modified formula from text-only language models to deliver remarkable outcomes with unequaled computational efficiency.
This new model generates text-to-images at a state-of-the-art rate while utilizing five times less computing power than earlier transformer-based techniques. It maintains low training costs and high inference efficiency while combining the adaptability and efficiency of autoregressive models. As a causal masked mixed-modal (CM3) model, CM3leon enhances the capabilities of prior models by being able to produce text and image sequences dependent on arbitrary sequences of other text and image content.
Meta and Microsoft Introduce the Next Generation of Llama
Meta and Microsoft jointly announced support for the Llama 2 family of large language models (LLMs) on Azure and Windows. Llama 2 aims to empower developers and organizations to create generative AI-powered tools and experiences. Microsoft Azure customers can now easily fine-tune and deploy the 7B, 13B, and 70B-parameter Llama 2 models, benefiting from Azure AI's powerful tooling for model training, fine-tuning, inference, and AI safety capabilities.
Additionally, Llama 2 will be optimized to run locally on Windows, making it convenient for Windows developers to integrate generative AI experiences into their applications using world-class tools like Windows Subsystem for Linux (WSL), Windows Terminal, Microsoft Visual Studio, and VS Code.
The inclusion of Llama 2 in the Azure AI model catalog expands Microsoft's AI offerings and provides developers and machine learning professionals with access to a hub of foundation models. This model catalog streamlines the process of discovering, evaluating, customizing, and deploying pre-built large AI models at scale, eliminating the need for managing infrastructure dependencies during the operationalization of Llama 2.
Elon Musk Launched xAI, His Own AI Startup
Elon Musk announced his latest startup, xAI, which will be focused on artificial intelligence. According to the company’s website, the goal of xAI is to “understand the true nature of the universe.” The xAI team will be led by Musk and others who have previously worked with OpenAI, DeepMind, Google Research, Microsoft Research, and Tesla.
Musk was an early backer of ChatGPT-creator OpenAI, but later criticized the company for inputting safeguards that aim to prevent the viral chatbot from spewing biased or sexist responses.