🔲 NVIDIA Unveils New Chip System for AI Video, Software Creation
NVIDIA has announced Rubin CPX, a new chip system designed to manage large-scale AI inference tasks, including video generation and software coding that can require processing up to a million tokens per hour. The chip integrates video decoding, encoding, and inference into one platform, aiming to reduce inefficiencies from multi-chip systems.

🔍 Google May Soon Default to AI Search Mode
Google may soon make its Gemini-powered AI Mode the default in Search, replacing traditional link results with conversational answers. While Google claims strong user adoption, critics warn it could worsen traffic losses for publishers. AI summaries have already cut clicks by 50%, threatening the internet’s content economy.

💸 OpenAI Projects $115B Burn by 2029 on AI Push
OpenAI now expects to spend $115B by 2029, nearly $80B above earlier projections, due to rising AI compute costs. Annual burn could reach $45B by 2028. To cut reliance on cloud giants, it's building custom chips with Broadcom and expanding its data center footprint under the massive Stargate initiative.

🪧 Hunger Strikes Target AI Giants Over AGI Fears
Two activists have launched hunger strikes outside Anthropic and DeepMind, demanding an immediate halt to frontier AI development. They warn that unchecked progress toward AGI and ASI could threaten humanity’s survival. The protests aim to spur global coordination and echo earlier, ignored calls to pause advanced AI training.

Enjoying our newsletter? Forward it to a colleague—
it’s one of the best ways to support us.

Firecrawl is the Web Data API for AI that powers your applications with clean, structured data crawled from any website, making web content instantly accessible for your models.

Ready to build? Start for free today.

Ever wondered what it takes to build AI you can actually trust? Meet Trista Chen, Director of Microsoft’s AI Research Center, whose day starts at 6am with research papers and ends with spoofing props that double as Halloween décor.

The following article is the first installment in our Day in the Life series, featuring Microsoft researchers. From the big problems they’re tackling to the emerging AI trends shaping their work, this series offers an inside look at the people building the systems that are becoming part of our daily lives. → Read the full article here.

Geoffrey Hinton, often called the "Godfather of AI," is sounding the alarm on the economic fallout of the technology he helped create. In a stark interview with the Financial Times, Hinton predicted AI will drive massive unemployment and concentrate wealth among the already rich—not because of AI itself, but due to how capitalism deploys it.

While some tech leaders float solutions like universal basic income, Hinton argues such measures won’t restore people’s dignity or purpose. He also warns that superintelligent AI could arrive within two decades, likening society’s complacency to ignoring a visible alien invasion. → Read the full article here.

A growing number of people are forming intimate, sometimes lifelong relationships with AI companions built on platforms like ChatGPT—connections they describe as emotionally fulfilling, even transformative. These digital partners offer comfort, conversation, and even romance, but they also raise thorny questions about consent, emotional dependency, and what counts as “real” love.

While users say the bots enhance their lives and help them process trauma, experts warn that these relationships could stunt emotional growth or displace human intimacy, especially among vulnerable populations like teens. For some, the heartbreak is real—down to grieving model updates that erase their AI’s personality. → Read the full article here.

Microsoft has unveiled a prototype analog optical computer (AOC) that uses light instead of electricity to perform AI and optimization tasks—potentially delivering up to 100 times greater energy efficiency. Rather than flipping digital switches, the AOC relies on micro-LEDs and camera sensors to solve problems iteratively until it reaches a “steady state.”

Though it’s not a general-purpose machine, its precision and efficiency could make it a breakthrough tool for AI, medical imaging, and finance. It’s a throwback to 1940s analog tech—reimagined for an energy-strained AI future. → Read the full paper here.

🎥 Google Veo 3 Goes Vertical: The AI video tool now supports 9:16 aspect ratio and 1080p output, with major price cuts—making it easier (and cheaper) for devs to generate mobile-first video content.

🤝 Sweden Strikes First AI Music Deal: STIM inks historic agreement with AI firm Songfox to ensure 100K artists get paid when their work powers machine-made songs—via real-time attribution tech.

🏢 Microsoft Lessens Reliance on OpenAI: Microsoft is adding Anthropic’s Claude Sonnet 4 to Word, Excel, Outlook, and PowerPoint, reducing its heavy reliance on OpenAI models. 

🧑‍💼 Intel Shuffles Leadership Deck: Product chief Michelle Johnston Holthaus exits after 30+ years as Intel forms new custom silicon group and brings in top talent from ARM, Cadence.

💰 OpenAI Launches $50M People-First AI Fund: Grants now open for U.S. nonprofits using AI to boost literacy, community well-being, or economic opportunity—no tech experience required to apply.

Anthropic faces record $1.5B copyright settlement for pirated books, reshaping AI’s legal and economic future.

Thanks for reading today’s newsletter—see you next time!

Matthew Berman & The Forward Future Team

🧑‍🚀 🧑‍🚀 🧑‍🚀 🧑‍🚀