🗞️ AI News Recap

Hey TEDchats Fam!

Happy Monday! Ever stumbled into a buzzing art gallery midway through an exhibition, surrounded by intricate masterpieces that you're desperate to comprehend? Navigating the world of AI news can feel a bit like that artistic exploration, with each piece contributing to the broader canvas of innovation. For today, consider us your gallery guide, ready to illuminate the mesmerizing details of the AI artworks showcased over the past week.

🐇 QUICK BITES — THE PAST WEEK

Reddit's Supposed $60 Million AI Licensing Deal Just Reached $200 Million!

Last week, a Bloomberg report revealed that Reddit had entered into a $60 million annual licensing deal with an unnamed major AI firm. This agreement allows the AI company access to Reddit's extensive archive of user-generated content, spanning over a decade, to train the unnamed company’s AI models. Now, Reddit's IPO prospectus, filed with the U.S. Securities and Exchange Commission, has revealed that aggregated contract deals signed in January 2024 amount to $203.0 million for terms ranging from two to three years (read Reddit’s official filing).

The House of Representatives Has a New AI Task Force

Leaders of the U.S. House of Representatives have introduced a bipartisan Task Force aimed at exploring ways Congress can maintain America's position as a global leader in AI innovation. To quote directly from the announcement, Mike Johnson stated, “It is important for Congress to work in a bipartisan manner to understand and plan for both the promises and complexities of this transformative technology”. Democratic Leader Hakeem Jeffries took it a step further, emphasizing that “Congress has a responsibility to facilitate the promising breakthroughs that artificial intelligence can bring to fruition and ensure that everyday Americans benefit from these advancements in an equitable manner” (see official press release).

Microsoft and Intel Forge a Billion-Dollar Bond

Microsoft and Intel, tech giants synonymous with the digital era, have formed a groundbreaking collaboration. At the Intel Foundry Direct Connect event, Intel unveiled plans to produce custom AI-era chips, with Microsoft as its largest customer in a $15 billion deal. Intel aims to regain chip manufacturing leadership, facing tough competition from companies like Nvidia and Apple relying on TSMC. This strategic move signifies a significant milestone for both companies, with Intel leveraging Microsoft's expertise in in-house designs for processors and AI accelerators. The collaboration, utilizing Intel's advanced 18A process, could mark a resurgence for Intel in chipmaking (read Intel’s announcement or see Intel’s announcement).

Stable Diffusion 3 Released for Advanced Text-to-Image Generation

Stable Diffusion 3 is now available in early preview, introducing an advanced text-to-image model that excels in multi-subject prompts, image quality, and spelling. The waitlist for early access is open, inviting individuals, developers, and enterprises to contribute insights for performance and safety enhancements before a broader release. With parameters ranging from 800M to 8B, the suite ensures scalability and quality options, aligning with the commitment to democratize access. Stable Diffusion 3 employs a diffusion transformer architecture and flow matching, with a detailed technical report set to be released soon (see official announcement).

Google's AI Faces Criticism for Rewriting Historical Figures

Google's Gemini AI faced backlash for generating historically inaccurate images of prominent figures, portraying the U.S. Founding Fathers and 1943 Nazi German soldiers as people of color. This controversial situation led Google to temporarily pause Gemini's image generation of people as the company works to refine its AI's understanding of historical accuracy (see Google’s full post).

Adobe’s New AI Assistant for PDFs

Adobe is taking a giant leap forward in digital document management with its new AI Assistant for Adobe Reader and Acrobat. This cutting-edge tool promises to revolutionize how we interact with lengthy documents. It's not just about reading anymore; the AI Assistant can suggest relevant questions, offer detailed answers, and even summarize entire documents. This means less time sifting through pages and more time gaining insights, making it a game-changer for students, professionals, and anyone dealing with extensive texts (see Adobe’s full announcement).

Bioptimus Raises $35 Million to Improve Public Health

Bioptimus, a new generative AI startup based in Paris, has raised $35 million in a seed funding round to focus exclusively on biology. The company, led by Jean-Philippe Vert, plans to tackle the unique challenges of working with sensitive, non-public clinical data to improve disease diagnosis and drug discovery. According to Vert, “By harnessing the power of foundation models and advanced algorithms trained on massive amounts of biological and multimodal data across scales, we aim to capture the laws of biology that have hitherto remained too complex to be properly understood”. Leveraging experts from Owkin and Google DeepMind, this initiative represents a significant investment in the intersection of AI and biological research (read more).

Europe’s Strong Stance Against the Tech Giants

In a world where a few tech giants dominate Artificial Intelligence, the European Union is taking a stand to ensure a fair AI landscape. The rise of generative AI, led by companies like Microsoft, Google, and Meta, has created an uneven playing field, prompting concerns about market monopolization. Margrethe Vestager, the EU's Executive Vice President for competition policy, stresses the need to scrutinize tech mergers, particularly those involving AI, due to the advantage these giants have with vast resources and partnerships. Vestager's assertive stance reflects a new approach in regulating the tech industry, suggesting potential stringent controls to level the playing field for smaller AI startups and prevent monopolistic practices (listen to the EU’s recorded discussion here).

Google Introduces Gemma for Responsible AI Development

Google introduces Gemma, a family of open models, continuing its commitment to accessible AI. Derived from Gemini, these models are lightweight, versatile, and designed for responsible AI development. Gemma offers model weights in two sizes, 2B and 7B, with pre-trained and instruction-tuned variants. A Responsible Generative AI Toolkit accompanies Gemma, aiding developers in creating safer applications. Gemma supports JAX, PyTorch, and TensorFlow frameworks, facilitating easy integration with tools like Hugging Face and NVIDIA NeMo (find out more here).

Sony Expands Its Virtual Reality (VR) Experiences

Sony is actively testing the PlayStation VR2 headset for use with PC gaming systems, a significant step beyond its current PS5 exclusivity. This development could enable PlayStation VR2 users to enjoy a wider variety of games on PC, potentially revolutionizing the VR gaming experience. Official support, along with additional game releases, is anticipated to be announced later in 2024 (see Sony’s announcement).

Chrome's New AI Tool for U.S. Users

With the launch of Chrome M122, users now have access to the "Help me write" feature on both Mac and Windows PCs in the United States. Powered by Gemini models, this feature aids in the creation and improvement of written content, whether for submitting a restaurant review, selling furniture, or making hotel reservation inquiries. According to Google, "the tool will understand the context of the webpage you’re on to suggest relevant content". For example, when drafting a review for garden shears, the tool can extract relevant information from a webpage, assisting in the writing process (see Google’s announcement and user guide).

Humane's Ai Pin Release Has Been Postponed

Humane's Ai Pin release, initially scheduled for March, has been postponed to mid-April, as announced by the company's Head of Media, Sam Sheffer. The first units are set to leave the factory by late March, with priority access customers receiving them in mid-April and remaining preorders shortly after. Founded by ex-Apple employees, Humane generated significant buzz before unveiling the Ai Pin in November. Positioned as a breakthrough in consumer hardware, the wearable represents a shift from smartphones and is backed by $230 million in funding. Preorders, priced at $699, include a three-month free subscription for orders before March 31 (see Humane’s announcement).

Unlocking Love's Potential with AI

Match Group, known for Tinder, Hinge, Match, and Plenty of Fish, has partnered with OpenAI, obtaining 1000+ ChatGPT Enterprise licenses. Match Group employees are utilizing these licenses for improved communication, coding, design, and analysis, sparking a shift in work culture. Will Wu, Match Group’s CTO, praises the impact in a recent press release, emphasizing AI as a productivity boost and innovation engine. However, responsible AI use is a priority and therefore, only trained and licensed Match Group’s employees will be allowed access (see Match Group’s press release).

🧰 AI TOOLBOX — RECAP

👋 UNTIL NEXT TIME

As we close the doors of the AI art gallery, each masterpiece of the past week has painted a vibrant canvas of technological innovation. The strokes of progress, the hues of breakthroughs, and the intricate details of advancements have created a tapestry that captivates the imagination. What future masterpieces await in the evolving landscape of AI? As we eagerly anticipate the next exhibition, join us for another journey through the world of AI in our upcoming newsletter.