AI This Week: Updates from ElevenLabs, OpenAI, NASA, and More

November 22, 2024

ElevenLabs, OpenAI, NASA, and Google have introduced new AI tools that transform how we interact with technology. ElevenLabs now offers the ability to create personalized conversational agents, while OpenAI’s integration of ChatGPT with macOS improves coding workflows. NASA’s collaboration with Microsoft has resulted in Earth Copilot, which simplifies access to geospatial data, and Google’s “Learn About” enhances the way users explore online content. Here’s everything you need to know.

ElevenLabs Expands Offering with Customizable Conversational AI Agents

ElevenLabs, known for its pioneering voice cloning and text-to-speech API, unveiled a new capability to build conversational AI agents. Announced on Monday, this feature is available on ElevenLabs’ developer platform, empowering users to craft conversational agents tailored to their needs.

The new platform simplifies the creation process, addressing challenges such as integrating knowledge bases and managing user interactions. Developers can now initiate their projects by selecting from several templates or starting afresh, choosing variables like the agent’s primary language, tone of voice, and even the AI model—Gemini, GPT, or Claude. The platform also allows users to set the creativity level of responses through the “temperature” setting and impose limits on token usage.

Further customization is supported by adding knowledge bases through files, URLs, or text blocks and integrating proprietary large language models (LLMs). ElevenLabs has made its SDK compatible across Python, JavaScript, React, and Swift platforms, with additional support via a WebSocket API for enhanced customization.

This move places ElevenLabs in direct competition with major tech giants and other AI startups, leveraging its unique position with customizable and flexible solutions that might edge out existing offerings from companies like OpenAI.

ChatGPT Bridges Desktop Apps for Enhanced Developer Experience

OpenAI has rolled out a new capability for the ChatGPT desktop app on macOS that allows it to interact directly with code from several developer-centric applications, such as VS Code, Xcode, TextEdit, Terminal, and iTerm2.

Until now, developers had to input their code into ChatGPT to utilize its capabilities manually. With the new “Work with Apps” feature, OpenAI simplifies this process by allowing the AI to access the code directly in a developer’s active window. This integration leverages macOS’s accessibility API to read text, ensuring that ChatGPT can analyze and respond to code without additional steps by the user. Although users still need to manually integrate ChatGPT’s output into their projects, this update significantly enhances the AI’s utility for coding tasks.

This development comes as OpenAI prepares for the anticipated launch of a general-purpose AI agent, codenamed “Operator,” which promises to integrate AI into daily computing tasks further.

NASA Introduces AI Earth Copilot to Simplify Access to Planetary Data

NASA has announced a new collaboration with Microsoft to launch an AI-driven chatbot, Earth Copilot, to enhance public access to its extensive scientific data about our planet. This initiative promises to transform how researchers and the general public interact with geospatial information, making it more digestible and more accessible to retrieve.

Earth Copilot is designed to field questions and provide clear, concise answers using NASA’s vast repository of Earth data. Users could inquire about the environmental impact of specific events and receive instant, straightforward explanations. For instance, they could ask about the effects of deforestation in the Amazon, the rate of glacial melt in polar regions, the impact of urbanization on local climates, or the consequences of increased plastic pollution in our oceans.

The project aims to “democratize” access to this critical scientific data, which has traditionally been challenging to navigate for non-experts. Tyler Bryson, Microsoft’s Corporate Vice President of Health and Public Sector Industries, highlighted that the tool aims to eliminate the complexity of data extraction and analysis, offering insights within seconds.

Currently, Earth Copilot is in its evaluation phase, accessible only to NASA scientists and researchers. They are testing the tool’s effectiveness on the Visualization, Exploration, and Data Analysis (VEDA) platform, which already provides access to a subset of NASA’s data. The forthcoming public release of this tool represents a significant advancement in making detailed planetary data accessible to a broad audience.

Google Unveils “Learn About”: A New Dimension in AI-Powered Search

Google has quietly rolled out a new AI-driven search experiment named “Learn About,” a part of Google Labs and its broader Learning Initiative. This tool redefines the search experience by summarizing content and offering a structured exploration path via interactive menus, enabling users to navigate related subtopics easily.

Key Features of Learn About

– Interactive Lists: This primary feature uses drill-down navigational menus that encourage exploratory learning. As users explore further, they uncover more layers of content, each connected to related subtopics.
– Conversational AI: “Learn About” is crafted to grasp new topics and deepen users’ understanding through a conversational, AI-powered interface that adapts to the queries posed.
– Visual Engagement: The tool enhances comprehension by pairing summaries with images sourced from stock providers like Shutterstock and Adobe, appealing to our visual orientation.
– Accessibility and Multilingual Support: Currently available exclusively to users over 18 in the United States and in English, “Learn About” intriguingly interacts with Spanish queries, indicating a potential for broader multilingual support in future updates.

Privacy and User Control

Google has implemented robust privacy controls within “Learn About.” Users must consent to a detailed agreement that outlines how their data is handled, including anonymization, before human reviewers analyze interactions for quality improvement.

Weekly Tool Highlight: Hexowatch

This week’s tool highlights features of Hexowatch, an AI-powered platform revolutionizing how businesses monitor websites. Trusted by over 130,000 companies, Hexowatch offers an array of monitoring types to keep you informed about any changes across your web pages at scale.

Key Features of Hexowatch

– Visual Monitoring: This feature enables monitoring for any visual changes across web pages, helping you spot both minor updates and major redesigns.
– HTML and Source Code Monitoring: This service tracks changes to specific HTML elements or modifications in your website’s source code.
– Keyword and Content Monitoring: Keeps an eye on specific keywords and visible content changes, ensuring your website stays up-to-date with your SEO strategy.
– Technology Monitoring: Alerts you to changes in your site’s tech stack or third-party scripts, keeping your technical operations smooth.
– Automatic AI Monitoring: Hexowatch autonomously detects all types of changes—visual, content, source code, and technological—ensuring comprehensive coverage.
– Availability and Domain WHOIS Monitoring: Guarantees that your site remains operational and tracks crucial domain registration details.

Applications Across Industries

– E-commerce: Stay ahead by monitoring competitor product launches and price changes, ensuring your offerings remain competitive.
– Real Estate: Get real-time alerts when properties that meet your criteria hit the market.
– Recruitment: Keep an eye on job openings at companies of interest or track updates from major job portals.
– Legal and Compliance: Maintain records of every site change with automatic archiving for legal purposes.