AI This Week
Leading AI companies like OpenAI, Google, and xAI are engaged in an intense competition for elite AI researchers, offering multi-million dollar compensation packages. The scarcity of top talent has led to creative hiring strategies and retention efforts. Some researchers are receiving bonuses of $1-2 million and equity increases of $20 million or more. Companies are also recruiting from diverse fields like theoretical physics and quantum computing. This talent war has intensified following recent high-profile departures from major AI labs.
Meta has postponed the release of its flagship AI model, Behemoth, originally slated for April 2025, now delayed to fall or later. Despite public claims of superiority over rivals, internal assessments reveal performance issues and training challenges. Engineers question whether the model's enhancements over previous versions justify a public launch. This delay raises concerns about Meta's substantial AI investments, with up to $72 billion in capital expenditures planned for the year. Internal tensions have surfaced, with executives reportedly blaming the Llama 4 team and considering management changes. The setback reflects broader industry challenges, as competitors like OpenAI and Anthropic also face delays in launching advanced models. Meta's stock fell 2.4% following the news.
TikTok has released AI Alive, a new tool that animates still photos for TikTok Stories, creating motion-filled video clips. This feature automatically transforms everyday images into dynamic visuals, adding sound and movement for a more engaging user experience. For example, a still photo of a sunset can evolve into a sweeping, cinematic video with atmospheric effects. Beyond its creative potential, the tool emphasizes safety by labeling AI-generated content and embedding metadata for transparency. AI Alive taps into advancing technology to explore new storytelling possibilities while aiming to maintain authenticity.
AI search startup Perplexity has skyrocketed to a $1.04 billion valuation after securing $62.7 million in new funding. The latest round, led by Daniel Gross and joined by investors including Jeff Bezos, Nvidia, and IVP, highlights growing interest in Perplexity’s real-time, answer-focused alternative to Google. With just 40 employees, the company is already generating millions in revenue and attracting 10 million monthly users. Perplexity’s AI offers cited answers with source links, distinguishing it from traditional search engines. The funds will be used to expand product development and scale operations. As the AI arms race intensifies, backers are betting on Perplexity to reshape how people find information online.
Perplexity has announced a partnership with PayPal to bring conversational commerce to life within its Perplexity Pro platform. Launching this summer in the U.S., this collaboration allows users to shop, book travel, and secure tickets instantly with PayPal or Venmo—all through Perplexity’s chat interface. Aravind Srinivas, CEO of Perplexity, emphasizes the shared focus on trust, while PayPal CEO Alex Chriss highlights the ease and security of shopping during chats. Key features include seamless payment, shipping coordination, tracking, and passkey-powered checkouts. This innovation also taps into PayPal’s global network of over 430 million accounts across 200 markets, all backed by strong security protocols.
Philips and Nvidia are working together to develop AI-driven advancements for MRI systems. The partnership focuses on creating a new AI model designed to improve the quality of MR images, reduce scanning times, and support more accurate diagnoses. Radiologists may find tools such as interactive image adjustments and automated scanning processes particularly relevant for enhancing workflow efficiency. This collaboration highlights ongoing efforts to modernize medical imaging and demonstrates how AI continues to play an important role in healthcare technology.
Apple is investigating the integration of AI-driven search functionality into its Safari browser, a move that could shake Google's dominance in online search. Eddy Cue, an Apple executive, revealed this potential shift during testimony in Google's antitrust trial with the U.S. Department of Justice. Analysts estimate that Google pays Apple roughly $20 billion annually to remain Safari’s default search engine, a lucrative arrangement now under scrutiny. With Apple’s ecosystem spanning 2.5 billion devices, any changes in search priorities could reshape the industry. Cue pointed to rising user interest in AI platforms like OpenAI and Perplexity, hinting at their future inclusion in Safari. Google’s ad business might face major repercussions.
Artificial intelligence is quickly shifting from an emerging trend to a foundational tool in the real estate industry. According to T3 Sixty’s latest Tech 200 list, more than half of the featured products now integrate AI, a notable rise from just under a third last year. Brokers and agents are adopting AI for tasks like lead management, content creation, and market analysis. Industry leaders are warming to the widespread use, with only 42% expressing significant concern — down from 50% a year ago. Alongside AI adoption, consolidation is reshaping the sector, with key acquisitions occurring among tech firms. Despite mergers, fresh innovators are making their mark, while established players dominate the landscape.
Groundbreaking research led by Yong Chen at the University of Pennsylvania's Perelman School of Medicine could revolutionize mental health diagnosis and treatment. Chen's work focuses on artificial intelligence that could detect mental health conditions — like dementia and depression — years before they fully manifest. Incorporating biometrics, psychometric testing, and smart-device data, the AI aims to revolutionize early diagnoses and personalized treatments. While still in the research phase, actionable insights are expected within two years, facing challenges such as privacy concerns and data variability. These tools promise to reshape mental health care, yet regulatory and technical hurdles remain.
Researchers from the University of Rochester and other institutions have developed MagicTime, an AI text-to-video model that generates realistic metamorphic videos. The system learns from time-lapse footage to simulate complex processes like plant growth or construction. MagicTime produces short, high-resolution clips that more accurately depict physical, chemical, and biological changes compared to previous models. This advancement moves AI closer to effectively modeling real-world transformations. The technology could potentially aid scientists in preliminary research by allowing faster exploration of ideas before conducting physical experiments.