March 10, 2026 News
Mira Murati's Thinking Machines Lab Secures Major Nvidia Compute Partnership for AI Development
Thinking Machines Lab, founded by former OpenAI co-founder Mira Murati, has signed a multi-year strategic partnership with Nvidia to deploy at least one gigawatt of Vera Rubin systems starting in 2027. The seed-stage company, valued at over $12 billion with $2 billion raised, is developing AI models that create reproducible results but has not yet released any products.
Skynet Chance (+0.01%): Massive compute scaling enables more powerful AI systems, but the focus on reproducible results could marginally improve control and reliability. The net effect is a slight increase in risk due to capability advancement outweighing the reliability focus.
Skynet Date (-1 days): The deployment of gigawatt-scale compute infrastructure accelerates the timeline for developing more capable AI systems that could pose control challenges. This represents significant acceleration in available resources for frontier AI development starting in 2027.
AGI Progress (+0.02%): A multi-billion dollar compute deal enabling gigawatt-scale deployments represents substantial progress in the infrastructure necessary for AGI development. The partnership between a well-funded AI lab and leading chip manufacturer signals serious commitment to advancing frontier AI capabilities.
AGI Date (-1 days): Securing gigawatt-scale compute starting in 2027 significantly accelerates the timeline for AGI by providing the computational resources needed for training increasingly capable models. This level of infrastructure investment suggests AGI development could proceed faster than scenarios without such massive compute availability.
Yann LeCun's AMI Labs Secures $1.03B to Develop World Models as Alternative to LLMs
AMI Labs, cofounded by Turing Prize winner Yann LeCun, has raised $1.03 billion at a $3.5 billion valuation to develop world models based on Joint Embedding Predictive Architecture (JEPA). Unlike traditional large language models, world models aim to learn from reality rather than just language, with initial applications planned in healthcare through partner Nabla. The ambitious project focuses on fundamental research and may take years before producing commercial applications, with the startup committing to open research and code sharing.
Skynet Chance (-0.03%): The focus on world models that understand reality through grounded learning and the emphasis on safety-critical applications like healthcare suggests a more controlled approach to AI development compared to less interpretable LLMs. The commitment to open research also enables broader safety scrutiny, though the fundamental capability advancement carries minimal inherent risk increase.
Skynet Date (+1 days): The multi-year fundamental research timeline and focus on safer, more grounded AI architectures rather than rapidly deployable products suggests a more deliberate development pace. This measured approach with extensive testing in real-world scenarios before deployment pushes potential risk timelines further out.
AGI Progress (+0.04%): World models that learn from reality rather than just language represent a significant architectural shift toward more general intelligence, addressing key LLM limitations like hallucinations and grounding. The substantial funding ($1.03B) and heavyweight team including LeCun, plus major backing from NVIDIA and other tech giants, indicates serious progress toward systems with broader understanding.
AGI Date (-1 days): The massive billion-dollar funding round, top-tier research talent, and major compute investment significantly accelerate the development of world models as a promising AGI pathway. Despite the multi-year timeline mentioned, the resource commitment and parallel efforts by competitors like Fei-Fei Li's World Labs suggest this approach is rapidly maturing toward AGI-relevant capabilities.