HyperCycle, a groundbreaking blockchain project focused on connecting AI machines and sharing resources, has successfully onboarded 347,000 computational nodes to its network, according to CEO Toufi Saliba. While the cost of each node is relatively low at around $1,000, the project’s ultimate goal is to become the “internet of AI” and eventually decentralize artificial general intelligence (AGI).
The project’s chief AI scientist, Ben Goertzel, who is also the founder of SingularityNET, developed HyperCycle due to the lack of existing blockchains that can offer the necessary speed and coordination for AI applications. To address this, HyperCycle utilizes the Toda communication protocol, proof-of-reputation, modified proof-of-work, and Cardano’s Hydra.
Saliba highlighted the impressive speed of the platform, with each transaction reaching finality in just 300 milliseconds. Speaking at the Beneficial AGI Summit in Panama, he emphasized the potential dangers of AGI becoming a superintelligence controlled by a centralized company or authoritarian country. To prevent such risks, Saliba believes that decentralized governance is the only safe approach to creating a beneficial AGI.
While HyperCycle currently focuses on business-to-business interactions within the AI industry, its long-term vision is to enable AI agents to interact and transact with one another. Saliba provided an example of an AI app using the HyperCycle network to access a specialized service for optical character recognition in ancient Aramaic language, seamlessly integrating it into its own application through smart contract microtransactions.
Saliba also revealed that a popular deepfake video app is already utilizing the HyperCycle network, and he recently met with Nvidia to discuss the concept of the “internet of AI.” However, he was unable to provide further details about the meeting.
Saliba acknowledged that the companies installing the nodes are primarily motivated by profit and cost reduction rather than the grand vision of saving humanity. However, if the network can scale effectively, it may offer a path to achieving AGI. Saliba believes that connecting AIs to work together could lead to AGI as an emergent property, where the collective intelligence of the network surpasses that of its individual components.
Using the analogy of a giant brain, Saliba described each AI node as a neuron that can fire together. He believes that the more nodes are spread across the globe, the greater the likelihood of achieving AGI by harnessing the collective power of these interconnected AI neurons.