🚀 NVIDIA launches Nemotron 3 Open Models
NVIDIA unveiled Nemotron 3, a new family of open AI models designed for efficient, agentic AI systems. The models use a hybrid mixture-of-experts architecture to boost throughput and cut inference costs. Nemotron 3 Nano is available now, with Super and Ultra coming in 2026. NVIDIA also released open datasets and reinforcement learning tools to support customization.

🔲 NVIDIA Weighs H200 China Production Surge
NVIDIA is considering ramping up production of its H200 AI chips after U.S. approval to sell them in China. Chinese firms are rushing to place orders following months of pent-up demand under export controls. The approval comes with a condition that a 25% fee be applied to H200 sales in China. Chinese regulators still must decide whether to allow imports.

💼 OpenAI Ends New-Hire Equity Vesting Cliff (Paywall)
OpenAI has removed its six-month equity vesting requirement for new employees. The change is meant to reduce risk for hires amid fierce competition for AI talent. Rivals including xAI have made similar moves as compensation packages soar. OpenAI expects to spend about $6 billion on stock-based pay this year, nearly half of projected revenue.

📉 Luminar Files For Bankruptcy
Luminar, once a leading lidar supplier for self-driving cars, has filed for bankruptcy following years of turmoil. The company plans to sell its lidar and semiconductor businesses while continuing limited operations during proceedings. Legal battles with ousted founder Austin Russell and multiple layoffs weakened the firm. Luminar was valued near $3 billion in 2020.

CompactifAI by Multiverse Computing compresses LLMs by up to 80% while maintaining accuracy, delivering up to 50% cost and energy savings for companies like IBM, Bank of Canada, and Moody’s. Run your models on fewer GPUs or even on the edge when you deploy CompactifAI to scale your AI workloads!

Try CompactifAI for Free and Realize Immediate Savings!

Disney partners with OpenAI on Sora, Adobe enters ChatGPT, Meta rethinks open source, DeepSeek chips drama, Runway Gen-4.5, and more AI news!

Universal basic income (UBI) has resurfaced in U.S. policy debates, led again by Andrew Yang, as fears grow that artificial intelligence could displace large segments of human labor. Columnist Eduardo Porter argues that while automation risks are real, UBI—as currently proposed—cannot address the scale of inequality and fiscal strain an AI-dominated economy would create.

A modest stipend, like Yang’s $1,000-per-month plan, would leave families below the poverty line, while a truly livable UBI would cost more than 40% of U.S. GDP. More fundamentally, Porter contends, UBI fails to confront the concentration of wealth and power among the owners of AI and capital, which—not unemployment alone—defines the deeper challenge. → Read the full article here.

The United States leads the world in AI competitiveness by a wide margin, according to Stanford University’s Global AI Vibrancy Tool, with China and India ranking second and third. The index, summarized by Visual Capitalist, measures AI strength across 42 indicators spanning research, investment, talent, policy, and public engagement—not just raw R&D output.

While the U.S. benefits from massive private investment and top research institutions, China’s surge in patents and deployments keeps it firmly in second place. Smaller, high-income countries like Singapore and the UAE outperform many larger economies relative to size, highlighting the role of targeted policy and governance strategies. → Read the full article here.

Rodney Brooks, the roboticist behind the Roomba and a foundational figure in modern robotics, argues that Silicon Valley’s push toward general-purpose humanoid robots is headed for failure. In a widely debated essay and interviews, Brooks says current technology cannot safely or reliably deliver human-level dexterity, balance, and judgment—despite billions flowing into startups promising do-everything robots.

He predicts a familiar cycle of hype followed by disappointment, with today’s humanoids largely forgotten within 15 years. The real risk, Brooks says, is that the obsession with humanoid forms diverts attention from more practical, specialized robots that could deliver real value sooner. → Read the full article here. (Paywall)

🚩 Grok Misreports Bondi Beach Shooting: Gizmodo found xAI’s chatbot misidentified the bystander who disarmed a gunman, questioned real footage, and cited false names.

🛡️ Google Engineer’s Four AI Safety Rules: An AI security engineer urges treating chatbots like a public postcard, deleting histories, and sticking to trusted platforms to reduce risks.

🤔 AI Mistakes Clarinet for Gun: A Florida middle school locked down after a weapon-detection system flagged a student’s clarinet, reviving concerns about oversight of AI surveillance.

💰 First Voyage Funds AI Habit Companion: The startup raised $2.5 million to grow Momo Self Care, an AI pet that nudges users toward habits with reminders and rewards.

🎬 Disney–OpenAI Deal Goes Wide in 2026: Disney’s integration of OpenAI tools like ChatGPT into parks and marketing is exclusive for one year. After 2025, other studios can partner too.

🔈 Want to advertise with Forward Future? Reach 600K+ AI enthusiasts, tech leaders, and decision-makers. Let’s talk—just reply to this email.

Thanks for reading today’s newsletter—see you next time!
Matthew Berman & The Forward Future Team

🧑‍🚀 🧑‍🚀 🧑‍🚀 🧑‍🚀