OpenAI Models See Sapient Outperformance While Google Boosts Gemini

A significant development in the AI world sees 22-year-old co-founders William Chen and Guan Wang of Sapient Intelligence rejecting a multimillion-dollar offer from Elon Musk's xAI. These friends from Michigan developed the Hierarchical Reasoning Model (HRM), a prototype with only 27 million parameters that, in June, outperformed larger systems from OpenAI, Anthropic, and DeepSeek on complex reasoning tasks like Sudoku and maze-solving. HRM's unique two-part design mimics human thought, leading to fewer 'hallucinations' than other AI models. Sapient Intelligence plans to open a US office soon. Meanwhile, the competition between Google and OpenAI continues to evolve. Google's AI, Gemini, now leads OpenAI's ChatGPT in user engagement, measured by the duration users spend on its web platforms per visit. This improvement follows recent updates, including Gemini 3 Pro and Nano Banana Pro. While ChatGPT, launched three years ago, still maintains a substantial lead in overall user numbers with over 800 million weekly users and 1.1 billion monthly web visits as of October, Gemini's app downloads grew significantly from 15 million to 66 million monthly by late 2025. Google's Pixel marketing VP, Adrienne Lofton, highlights Gemini's integration into Google Suite and camera features like Auto Best Take, aiming to show clear benefits to consumers. The broader AI industry is also seeing diverse applications and future plans. Rubrik launched Rubrik Agent Rewind in November 2025, a new platform for AI agent security, with integrations for Microsoft Copilot Studio and enhanced protections for Microsoft 365, Azure DevOps, and GitHub. CES 2026 in Las Vegas will showcase AI's pervasive integration across products, featuring a keynote by AMD's Lisa Su and discussions on topics from cinematic creativity to industrial robotics. The banking industry is moving towards 'agentic AI,' with a McKinsey & Company report predicting potential cost cuts of 15-20% by 2030, though it also risks $170 billion in lost profits if banks do not adapt. Experts like Jonathan Pelosi from Anthropic believe 2026 will be a turning point for AI accuracy in this regulated sector. Ethical considerations remain central, with the Mint All About AI Tech4Good Summit 2025 emphasizing AI's positive societal impact and the need for responsible use, while newsrooms adopt AI tools for efficiency with strict ethical rules, ensuring a 'human in the loop' for fact-checking.

Key Takeaways

• William Chen and Guan Wang of Sapient Intelligence rejected a multimillion-dollar offer from Elon Musk's xAI.
• Sapient Intelligence's Hierarchical Reasoning Model (HRM) prototype, with 27 million parameters, outperformed larger models from OpenAI, Anthropic, and DeepSeek in June on complex reasoning tasks.
• Google's Gemini now leads OpenAI's ChatGPT in user engagement per web visit, following updates like Gemini 3 Pro and Nano Banana Pro.
• ChatGPT still holds a significant lead in overall user numbers, with over 800 million weekly users and 1.1 billion monthly web visits as of October.
• Rubrik launched Rubrik Agent Rewind in November 2025 for AI agent security, integrating with Microsoft Copilot Studio and enhancing protections for Microsoft 365, Azure DevOps, and GitHub.
• CES 2026 will focus on AI integration across products, featuring a keynote by AMD's Lisa Su.
• The banking industry is adopting 'agentic AI,' with McKinsey & Company projecting 15-20% cost reductions but also a risk of $170 billion in lost profits by 2030 if banks fail to adapt.
• An ad agency, Episode Four, uses AI 'hallucinations' from its RYA tool to generate unique advertising ideas, cutting ideation time from weeks to days.
• Newsrooms are using AI tools for efficiency, such as transcribing interviews and brainstorming headlines, but maintain strict ethical rules requiring human oversight for fact-checking and content creation.
• The Mint All About AI Tech4Good Summit 2025 highlighted AI's potential for public good and the importance of ethical, responsible AI development, including India's need for 'sovereign AI' models.

Young Innovators Reject Elon Musk Offer for AI Breakthrough

William Chen and Guan Wang, 22-year-old friends from Michigan, turned down a multimillion-dollar offer from Elon Musk's xAI. They are co-founders of Sapient Intelligence and developed the Hierarchical Reasoning Model (HRM). In June, their HRM prototype, with only 27 million parameters, outperformed larger systems from OpenAI, Anthropic, and DeepSeek on complex reasoning tasks. HRM uses a unique two-part design that mimics human thought, leading to less "hallucination" than other AI models. The company plans to open a US office soon.

Michigan Friends Reject Musk Offer to Build New AI

William Chen and Guan Wang, founders of Sapient Intelligence, are making big news in AI. These childhood friends from Michigan turned down a multimillion-dollar offer from Elon Musk's xAI. They created the Hierarchical Reasoning Model (HRM), an AI that excels in abstract reasoning. In June, their HRM prototype, with just 27 million parameters, beat larger systems from OpenAI, Anthropic, and DeepSeek. HRM uses a special two-part design to think like humans, reducing errors. Sapient Intelligence plans to open a US office soon.

Young Founders Reject Musk Offer Their AI Beats OpenAI

William Chen and Guan Wang, 22-year-old co-founders of Sapient Intelligence, rejected a multimillion-dollar offer from Elon Musk's xAI. These friends from Michigan developed the Hierarchical Reasoning Model (HRM). In June, their HRM prototype, with only 27 million parameters, outperformed systems from OpenAI, Anthropic, and DeepSeek. It excelled in complex tasks like Sudoku and maze-solving. HRM uses a unique two-part design to mimic human thinking, which helps it "think" instead of just guessing. Sapient Intelligence plans to open a US office soon.

Google Gemini Surpasses ChatGPT in User Engagement

Google's AI, Gemini, now leads OpenAI's ChatGPT in how long users spend on its web platforms per visit. This marks a big step for Gemini, which struggled with engagement in mid-2024. Recent updates, including Gemini 3 Pro and Nano Banana Pro, have made users spend more time with the AI. While ChatGPT still has more app downloads, Gemini's downloads grew from 15 million to 66 million monthly by late 2025. This shows Google's efforts to integrate Gemini into its products like Android and Google Workspace are working.

ChatGPT Keeps Lead Over Google Gemini in User Numbers

ChatGPT still holds a strong lead over Google's Gemini in user numbers, even though Gemini received a "smarter" update in November. ChatGPT, launched three years ago, now has over 800 million weekly users and 1.1 billion monthly web visits as of October. In contrast, Gemini had 153 million monthly web visits in October. OpenAI's CEO Sam Altman focuses on user engagement, adding features like "persistent memory" and "AI agents" to keep users coming back. Google's AI chief Demis Hassabis prioritizes scientific advances, and Google has historically struggled to build platforms with strong user "network effects."

Pixel Marketing VP Adrienne Lofton Discusses AI Strategy

Adrienne Lofton, Google Pixel's marketing VP, shared insights on the AI device market with Business Insider. She was once an iPhone user but now believes Pixel could become the top AI device. Lofton explained that Google integrates Gemini into its Google Suite to boost productivity and creativity. She noted that while many companies promote AI, 50% of consumers remain unsure about it. Pixel's marketing focuses on showing clear benefits, like how Gemini powers camera features such as Auto Best Take, rather than just talking about AI technology itself. Lofton believes the industry is moving faster than consumer demand for new AI-first devices.

Rubrik Launches New AI Agent Security Platform

In November 2025, Rubrik launched Rubrik Agent Rewind, a new platform for AI agent security. This technology helps monitor, control, and reverse unwanted actions by AI agents in businesses. Rubrik also announced integrations with Microsoft Copilot Studio and improved protections for Microsoft 365, Azure DevOps, and GitHub users. The company used artist Ludacris in its marketing to highlight its focus on leading in AI security. Rubrik aims to simplify and protect how businesses use AI agents, despite facing challenges like achieving product-market fit and current financial losses.

CES 2026 to Showcase AI Innovations in Las Vegas

CES 2026 will take place in Las Vegas in January, focusing on how AI is becoming a core part of almost every product. Companies like LG and Samsung will share their visions for future homes, while Lenovo will host a keynote at The Sphere. Sony Honda Mobility will feature its Afeela electric vehicle, and Waymo and Zoox will showcase robotaxis. The event includes many AI-focused panels, such as "The Future is Airborne" and "Transforming Industries with Physical AI." AMD's Lisa Su will deliver a keynote, and discussions will cover topics from cinematic creativity to industrial robotics.

Ad Agency Uses AI Hallucinations for Creative Ideas

Leslie Walsh, head of product strategy at Episode Four, a New York ad agency, uses AI "hallucinations" to create fresh advertising ideas for brands. Her team developed RYA, an AI tool that generates "radical ideas" based on consumer data from weekly surveys. They intentionally prompt the AI to produce unusual combinations and "nonsensical" outputs to spark unique concepts. This process has cut down ideation time from weeks to days, helping clients in financial services, auto, and travel. For example, they turned financial education into a celebrity chef series and an auto lending service into a dating show.

AI Predicts NFL Week 13 Game Outcomes and Best Bets

A self-learning AI from SportsLine's Data Science team is providing NFL Week 13 picks and score predictions. This AI uses advanced machine learning to analyze historical data and evaluate team strengths. For Week 13, the AI predicts the Rams will comfortably cover their -10.5 spread against the Panthers, forecasting a 30-17 win. Rams quarterback Matthew Stafford has performed exceptionally, and wide receiver Puka Nacua has been a key target. The AI PickBot has successfully made over 2,000 top-rated prop picks since the 2023 season began.

Mint Summit Highlights AI's Positive Impact and Ethics

The Mint All About AI Tech4Good Summit 2025, held in Mumbai on November 28, focused on AI's positive impact for society. Experts discussed using AI for public good, including sustainability, healthcare, and education. Binoy Prabhakar from HT Digital emphasized AI's role in enhancing human potential, not replacing it. Salesforce's Arundhati Bhattacharya spoke about ethical technology use. Amitabh Kant, former G20 Sherpa, urged India to develop its own "sovereign AI" models to ensure national security and economic independence. Discussions also covered generative AI's role in learning and the importance of responsible AI use.

Newsrooms Use AI Tools with Strict Ethical Rules

Newsrooms in New Hampshire and across the country are using AI tools to improve efficiency while setting clear ethical rules. Journalists use AI for tasks like transcribing interviews with Otter, brainstorming headlines, and converting PDFs to searchable documents. However, editors like Jonathan Van Fleet of the Concord Monitor and Julie Hirshan Hart of the Laconia Daily Sun stress that AI will never write articles or photos. They emphasize that a "human in the loop" must always fact-check and vet any AI-generated information. Newsrooms are also being transparent about their AI policies to build reader trust.

Banks Will Use AI with Precision in 2026

The banking industry is moving towards "agentic AI," which involves autonomous systems that can perform complex tasks with little human help. A McKinsey & Company report predicts this shift could cut bank costs by 15-20% but also risks losing $170 billion in profits by 2030 if banks do not adapt. Experts like Jonathan Pelosi from Anthropic believe 2026 will be a turning point as AI accuracy improves, closing the "trust gap" in this regulated industry. Banks are now focusing on specific business goals rather than just experimenting with AI. Key applications include updating old COBOL-based systems and automating compliance tasks like KYC.

Sources

• Two Young Innovators Reject Elon Musk’s Multimillion-Dollar Offer To Build New AI Breakthrough
• Michigan Innovators William Chen And Guan Wang Lead AI Revolution With Sapient Intelligence
• Gen Z Students Reject Elon Musk's Offer, Their AI Model Outperforms OpenAI
• Gemini Goes Past ChatGPT In Time Per Visit That Users Spend On The Platform
• Why ChatGPT Still Has a Winning Edge Over Google's Smarter AI
• Google Pixel's marketing VP says she was once a lifelong iPhone user
• The Bull Case For Rubrik (RBRK) Could Change Following Launch of AI Agent Security Platform
• The Ultimate CES 2026 Guide: Hottest AI Panels, Parties And Robots
• I lead product strategy at a creative agency and use AI 'hallucinations' to come up with ideas for brands
• Week 13 NFL picks, score predictions, best bets today from self-learning AI
• AI’s positive impact in focus at Mint All About AI Tech4Good Summit 2025
• AI in journalism: A balance of efficiency and ethics - Monadnock Ledger-Transcript
• Precision, Not Hype, Will Shape Banks’ Use Of AI In 2026