Global Supercomputer Network Aims to Accelerate AGI Development
Global, Monday, 12 August 2024.
Scientists are creating a worldwide network of supercomputers to advance artificial general intelligence (AGI) research. The first node, set to be operational by September 2024, marks a significant step towards human-level AI capabilities. This ambitious project could revolutionize AI development and impact various industries globally.
Ambitious Vision for AGI
The new supercomputing network, spearheaded by SingularityNET, aims to create a ‘multi-level cognitive computing network’ capable of training AGI architectures. This initiative is set to transition AI from big data learning to ‘non-imitative machine thinking’ and dynamic world modeling. The supercomputers will be equipped with cutting-edge components like Nvidia L40S GPUs, AMD Instinct processors, and Tenstorrent Wormhole server racks with Nvidia H200 GPUs[1].
Technological Integration and Components
SingularityNET’s modular supercomputer will integrate advanced AI systems, including deep neural networks, large language models (LLMs), and multimodal systems. These systems will connect human behaviors such as speech and movement inputs with multimedia outputs. The completion of this network is expected by late 2024 or early 2025, depending on supplier timelines. The system’s architecture will be supported by OpenCog Hyperon, an open-source software framework designed for AGI[2].
Operational and Strategic Insights
Ben Goertzel, CEO of SingularityNET, emphasizes that this supercomputer will be a major breakthrough in advancing AGI. The network aims to facilitate the growth of artificial superintelligence by providing access to data via a tokenized system, allowing users to contribute and utilize existing datasets for AGI testing. This decentralized approach is expected to enhance the speed and efficiency of computational tasks[3].
Global Impact and Future Prospects
The development of this supercomputing network is poised to revolutionize AI research and development, impacting various industries globally. It aims to surpass human intelligence across multiple disciplines and self-improve through continuous data access. Predictions from industry leaders like DeepMind’s Shane Legg and Facebook’s Mark Zuckerberg suggest that AI systems could match or exceed human intelligence by 2027-2028[1].
Location and Key Players
The project is primarily driven by SingularityNET, a decentralized AI company based in the Netherlands. Key figures include CEO Ben Goertzel and partners from the Artificial Super Intelligence Alliance, such as Fetch.ai. The company has recently invested $53 million to develop the world’s first modular supercomputer dedicated to AGI and Artificial Superintelligence (ASI) research and development. This investment underscores SingularityNET’s leadership in decentralized AI[4].