Researchers at Meta and the University of Southern California have developed Megalodon, a new architecture for AI models. It allows language models to process significantly larger amounts of text without using a lot of memory.
A mysterious chatbot called “gpt2-chatbot” caused a stir among experts this week. Its capabilities appeared to be on a par with GPT-4, but its origin remained unknown. After a short time, the bot disappeared from the scene again, presumably due to the high level of interest. Speculation about the developer, including OpenAI, Google and Anthropic, …
OpenVoice allows users to realistically clone voices in different languages and accents, and even control emotions and speaking styles. The latest version, OpenVoice V2, offers improved audio quality, native support for multiple languages, and is available free for commercial use. Source: Hacker News
Salesforce Einstein Copilot, an AI-powered tool for enterprise, is now generally available and is designed to help sales teams be more productive through generated text and automated actions. Einstein Copilot can break down and execute complex tasks, such as identifying the best sales opportunities and creating email drafts.
Cohere releases a toolkit to accelerate the development of generative AI applications in the enterprise. With pre-built applications and easy-to-implement components, the toolkit aims to reduce development time.
Snowflake introduces Arctic, a new open language model designed specifically for complex enterprise tasks such as generating SQL queries and code or following instructions.
Anthropic, creator of ChatGPT competitor Claude, introduces the “Team” business solution. It includes access to the three latest Claude models, has increased usage limits, admin tools and invoice management, and offers longer context windows for uploading large documents for editing. In addition, an iPhone app is now available.
The Amazon Q AI chatbot is now generally available to help businesses with tasks such as knowledge discovery, software development, and data analysis.
Apple releases OpenELM, a set of small, freely available AI models that can run directly on devices like laptops or smartphones and perform tasks such as text generation efficiently. While not industry-leading in performance, OpenELM seems to provide a solid foundation for future research and development in on-device AI.
Adobe integrates its Firefly AI image generator directly into Photoshop, allowing users to create images using text input and then edit them using familiar Photoshop tools.