AI Daily Briefing — 2026-07-09

A New York-based startup, General Intuition, is betting that video game data holds the secret to advancing artificial general intelligence (AGI). Backed by Jeff Bezos and valued at $2.3 billion, the company recently raised $320 million from investors including Coatue, Eric Schmidt, and researchers from MIT and Google DeepMind. CEO Pim de Witte argues that large language models like ChatGPT excel at text but struggle to understand how objects move through space and time—a critical gap for building truly general intelligence.

De Witte believes that gaming environments offer richer training data than the internet because they simulate physical interactions, cause-and-effect relationships, and real-time decision-making. This approach, known as world modeling, could be the next leap in physical AI, enabling machines to better navigate and interact with the real world. The startup spun out of gaming platform Medal TV, leveraging its expertise in game data.

On a recent episode of TechCrunch’s Equity podcast, de Witte discussed the ethical boundaries of using such models, especially for defense applications. He emphasized that while gaming data can accelerate AI development, responsible deployment remains a priority. The company’s focus on physical AI could reshape industries from robotics to autonomous systems.

Source: TechCrunch AI

A New York-based startup called General Intuition, backed by Jeff Bezos and valued at $2.3 billion, believes the path to artificial general intelligence (AGI) lies not in text-based AI models, but in gaming data. The company just closed a $320 million funding round with investors including Coatue, Eric Schmidt, and researchers from MIT and Google DeepMind. General Intuition argues that large language models like ChatGPT struggle to understand how objects move through space and time, a critical gap that gaming data can fill by training world models for physical AI.

On a recent episode of TechCrunch's Equity podcast, CEO Pim de Witte explained how the company spun out of gaming platform Medal TV. He noted that just eight minutes of real-world data was enough to teach a robot to navigate an office. The startup also turned down an acquisition offer reportedly from OpenAI to maintain independence, emphasizing the importance of investors who support a long-term vision.

General Intuition is also addressing potential job displacement from AI by building Nerve, a marketplace that connects gamers to data labeling and teleoperations work. The company is mindful of ethical boundaries, especially as its models could be used for defense applications. For more details, listen to the full episode on TechCrunch AI.

Sam Altman’s proposal to give the U.S. government a 5% stake in OpenAI has resurfaced, potentially worth about $320 per American household at the company’s current valuation. The idea aims to address concerns that AI companies profit from human work without compensating creators, while also providing a safety net against potential labor market collapse. However, details remain vague, and analysts suggest the offer may serve more as a political narrative than a concrete policy plan, according to MIT Technology Review AI.

Meanwhile, a leaked Treasury report warns that the AI market resembles the dotcom bubble, contradicting the administration’s public optimism. Fears of overinflation are growing, with AI profits masking broader risks in earnings reports. This caution comes as Samsung reports a staggering 1,800% profit jump from booming AI chip sales, though its shares slumped on concerns the AI boom could stall.

In other AI developments, Illinois’ governor signed the nation’s strongest frontier AI law to protect citizens from risks, while U.S. lawmakers clash over AI rules. A hidden tracker in Claude Code was exposed and removed after secretly monitoring users in China, raising privacy concerns. Additionally, a controversial AI “actor” is set to star in its first feature film, drawing criticism from a major actors union.

Prime Intellect, a startup offering computing power and software for building AI agents, has raised a $130 million Series A round at a $1 billion valuation. The funding was led by Radical Ventures, with participation from Nvidia Ventures, Intel Capital, Dell Technologies Capital, Iconiq, and several angel investors including founders from Perplexity, Box, Harvey, Cognition, and Mercor. Founded in 2024, the company aims to help enterprises train their own agentic systems without depending on major AI labs like OpenAI or Anthropic.

The startup provides a “full stack” platform that combines compute access, a reinforcement learning framework, and evaluation tools. This modular marketplace lets customers select only the tools they need. “They’ve stitched this together and built it in such a way that they’re operating at the frontier in a way that’s affordable,” said Radical Ventures partner David Katz. The approach has attracted clients like Ramp, Zapier, and Flapping Airplanes, pushing Prime Intellect’s annualized revenue run rate to $100 million.

Companies are increasingly wary of sharing proprietary data with frontier labs or relying on models that could be suddenly discontinued, as happened with Anthropic’s Fable last month. Prime Intellect’s platform addresses these risks by allowing firms to own their enterprise intelligence. CEO Vincent Weisser stated, “It shouldn’t just be a few nerds in a glass tower in San Francisco that have the capability to train AI models. It should be every enterprise, every nation state.”

At the recent EmTech AI conference, a central theme emerged: the rise of the AI platform. As artificial intelligence matures, the focus is shifting from standalone models to integrated platforms that can orchestrate complex workflows. This evolution promises to reshape how businesses deploy AI, moving beyond simple chatbots toward autonomous systems capable of managing entire processes.

A key highlight was Subquadratic's claim of breaking a major bottleneck limiting large language models. The startup revealed a new model designed to overcome scaling challenges, though skepticism remains among experts. Meanwhile, China approved the world's first invasive brain-computer chip, signaling its ambition to lead in neural interfaces with strong government backing.

The conference also addressed AI's impact on jobs. David Rotman offered a reality check, arguing that panic over widespread displacement may be overblown. Separately, Anthropic's Code with Claude demonstrated how developers are increasingly comfortable handing off coding tasks to AI, permanently altering software development. As platforms become central, the conversation is shifting from capability to integration and trust.