🗞️ AI News Recap

Hey TEDchats Fam!

Happy Monday! Ready for a journey through the latest AI buzz? This week's newsletter is jam-packed with thrilling updates from the past week! From a series of OpenAI news stories to ElevenLabs' innovative model, and even some surprising revelations from Hugging Face, there's excitement in abundance. Plus, we'll touch on the impact of AI demand on government decisions. Get ready to dive in!

🐇 QUICK BITES — THE PAST WEEK

OpenAI's Eventful Week: 8 Major AI Updates and News Stories!

• Free ChatGPT users can now access a suite of advanced features previously reserved for paid subscribers, including custom GPTs, data analytics, chart creation, and more, albeit with some limitations (read more).

• ChatGPT MacOS desktop app is now available to Plus and Team users, with a broader rollout planned for the coming weeks and a Windows version set to launch later this year (read more).

• ChatGPT Edu has launched and is tailored for universities to responsibly deploy AI to students, faculty, researchers, and campus operations. Powered by GPT-4o, it supports reasoning across text and vision, alongside advanced tools like data analysis (read more).

• OpenAI for Nonprofits has been announced, offering discounted access to ChatGPT. This includes ChatGPT Team at $20 per month per user, and larger nonprofits can receive a 50% discount on ChatGPT Enterprise (read more).

• OpenAI has published a ‘public threats’ report detailing their effective interventions in thwarting covert influence operations aiming to utilize their models for deceitful purposes across the internet (read more).

• Insights on CEO Sam Altman's 2023 firing was revealed, but not without controversy. In 2023, OpenAI's board ousted co-founder and CEO Sam Altman from the company. Despite Altman being reinstated, the exact reasons behind the decision remained unclear until now. Former board member Helen Toner provided insights, highlighting issues such as Altman's failure to disclose ownership of the OpenAI Startup Fund and providing inaccurate information about safety processes. Bret Taylor, the current OpenAI board chair, responded by expressing disappointment in Toner's remarks and reaffirming the board's commitment to the company's mission (read more).

• OpenAI has formed a new safety team headed by CEO Sam Altman and board members Adam D’Angelo and Nicole Seligman. The team’s mission is to enhance safety and security measures for OpenAI's projects. However, concerns have been raised about the team's effectiveness, especially since all three leaders hold positions on OpenAI's board. This move follows the resignation of several key AI researchers who cited safety concerns, including the departure of Superalignment team co-leaders Ilya Sutskever and Jan Leike (see OpenAI’s announcement).

• OpenAI’s competitor Anthropic recently hired an ex-OpenAI safety leader. With the AI field evolving rapidly, Jan Leike, who recently left OpenAI, has now joined Anthropic to lead a new 'Superalignment' team. This team will focus on scalable oversight, weak-to-strong generalization, and automated alignment research, with the aim of enhancing AI safety and security (see Leike’s announcement).

Introducing ElevenLabs' Revolutionary Text-to-Sound Model

This week, ElevenLabs stirred up excitement with the release of their groundbreaking text-to-sound model. Their newest AI audio technology is capable of producing sound effects, short instrumental tracks, immersive soundscapes, and an extensive range of character voices, all generated from a simple text input. Now accessible to all users, you can view their demo video for a closer look (check it out here).

Hugging Face Discloses AI Data Breach!

Hugging Face, a platform for creating and hosting AI models and related resources, disclosed a data breach on their Spaces platform on Friday, particularly concerning Spaces secrets, prompting swift action. Spaces secrets, essential for unlocking protected resources such as accounts and tools, were compromised. Immediate measures were taken, including the revocation of affected tokens and a request for users to update keys or switch to fine-grained access tokens. The breach is actively under investigation with cybersecurity experts, with a concerted effort to enhance security measures and phase out "classic" tokens, and authorities have been notified (see Hugging Face’s announcement).

With $6 Billion in Funding, Elon Musk Promotes Job Opportunities for His AI Startup

Elon Musk’s xAI recently secured $6 billion, with an $18 billion pre-money valuation, propelling it into the spotlight and intensifying its race against AI giants like OpenAI and Google. Musk posted on X (formerly Twitter) career opportunities for those interested in “understanding the universe”, emphasizing a rigorous pursuit of truth without regard for popularity or political correctness (see Musk’s post on careers).

Microsoft’s Copilot Bot Lands on Telegram

Microsoft has unveiled an official Copilot bot on the messaging platform Telegram. The update enables users to search, pose inquiries, and interact with the AI chatbot. Copilot for Telegram is currently in beta and can be accessed for free on both mobile and desktop devices. This move by Microsoft aligns with the trend among AI companies to make their language models (LLMs) available through messaging apps. This trend is also seen with Meta AI now accessible in Messenger, WhatsApp, and Instagram messaging. Additionally, Google offers similar features in their chat applications (see Microsoft’s announcement).

New AI-Powered Voice Search on Fire TV

Amazon is enhancing Fire TV with AI-powered voice search capabilities that allow Alexa to handle open-ended questions about movies and TV shows. This new feature will roll out in a few weeks to U.S. residents with Fire TV devices running Fire OS 6 or later. Soon, users will be able to ask their TV questions like 'What movie has the line, "You're killing me, Smalls?"' or request tailored recommendations based on their preferences and streaming subscriptions. This marks another step in Amazon’s efforts to integrate generative AI into its devices, enhancing the viewing experience with personalized and intuitive features (see Amazon’s announcement).

Alo Steps into Mixed Reality on Meta Quest 3

Experience a new level of immersive fitness with Alo Moves’ 3D classes on Meta Quest 3, where lifelike instructors guide you through yoga, pilates, and meditation in stunning virtual environments. Launching later this year, it will be complete with movable “mini-instructors” so you can easily see poses or movements from any angle (see Meta’s announcement).

Anthropic Unleashes Custom AI Assistants

Anthropic’s new feature for its AI chatbot Claude now enables users to create personalized AI assistants. Dubbed "tool use", this feature allows these bots to interact with any external API, offering functions like email management, shopping assistance, or even interior design consultation based on image analysis. Accessible through Anthropic Messages API, Amazon Bedrock, and Google Vertex AI, this tool is versatile and relatively affordable, with pricing based on text volume processed. This advancement represents a major progression in the evolution of AI, akin to OpenAI’s GPTs, rendering it more accessible and adaptable for a diverse array of applications (see Anthropic’s announcements here and here).

Perplexity AI Launches a ‘Shareable’ Feature

Perplexity AI introduces Perplexity Pages, enabling users to transform search queries into detailed, visually appealing, and shareable web pages (see Perplexity’s announcement).

Capture 3D Product Videos on Your iPhone

French startup AniML's new app, Doly, revolutionizes product video creation by letting users generate 3D models with their iPhones. Utilizing advanced AI techniques like Gaussian splatting, Doly transforms still images into realistic 3D models, making professional-grade product videos accessible to e-commerce entrepreneurs without the need for expensive equipment or studios. Users simply move their phone around the product to capture it in 3D, then choose from various templates to create stunning videos. This innovation democratizes high-quality visual content, allowing small businesses to compete with larger brands in the digital marketplace (see more).

AI Tutors Transform Traditional Learning

AI-powered tutoring apps are revolutionizing how students in the U.S. approach their studies. Leading the charge are Chinese-developed apps like Answer AI and Question AI, which offer personalized and affordable alternatives to expensive in-person tutoring. However, while these AI tutors enhance learning accessibility, they still face challenges with accuracy and reliability. As AI continues to evolve, its role in education remains an ongoing debate among educators and parents. While the debate continues, U.S. Senator Maria Cantwell has now proposed a bill to boost AI education, which would include AI guidance for K-12 teachers and new AI education hubs at community colleges. You can find more information on the matter here, here, and here.

Microsoft Challenges MacBook Air Dominance

Microsoft is boldly claiming its new Copilot Plus PCs, powered by Qualcomm’s Snapdragon X Elite, outshine Apple’s MacBook Air M3. Highlights include a 50% better performance in Cinebench 2024 multithreaded workloads and a 16% higher Geekbench 6 multithreaded score. In real-world tasks, the Surface Laptop encoded a 4K video faster than the MacBook Air M3, clocking in at just over five minutes. Moreover, Microsoft's new laptop excels in battery life, lasting nearly 17 hours on web browsing compared to the MacBook Air's 15.5 hours and delivering over 20 hours of video playback. With these benchmarks, Microsoft sets the stage for a competitive showdown in the ultrabook market, positioning the Surface Laptop as a formidable contender against Apple's bestseller (see Microsoft’s review here and The Verge’s review here).

AI Demand Forces Biden Administration to Reconsider Coal Phase-Out Plans

The Biden administration is contemplating delaying its plans to reduce coal power reliance due to the increased demand driven by AI technologies, as reported by the Financial Times. Although accurately quantifying AI's energy consumption remains challenging, there is a growing acknowledgment of its significant energy requirements. Concerns are mounting over grid stability and ensuring adequate power supply to AI data centers, especially if the U.S. intends to remain competitive in the global AI arena. What's evident is that as the U.S. endeavors to lead in artificial intelligence and manufacturing, the surge in power demand presents a challenge to its decarbonization targets (read more).

🧰 AI TOOLBOX — RECAP

👋 UNTIL NEXT TIME

As we wrap up our dive into the latest AI news from the past week, we've covered a wide range of stories, from OpenAI, ElevenLabs, and Hugging Face to many others. Each story sheds light on the evolving AI landscape. Stay tuned for further updates as we journey through the forefront of AI innovation together. Until next time, happy exploring!