The Revolution of General Artificial Intelligence Could Arrive Sooner Thanks to This Network of Supercomputers
General Artificial Intelligence (GAI) is considered the future holy grail of the AI industry. Unlike generative AIs that rely on large language models (LLMs), GAI will be capable of learning human-specific tasks in a cognitive manner. Currently, we are only in the early stages, and companies in the field are mostly focusing on generative AI, which, according to one of the leading figures in AI, Yann LeCun, doesn't have a sustainable future.
However, according to SingularityNET, GAI might soon take its first steps thanks to its new network of supercomputers. The company's CEO, Ben Goertzel, stated to LiveScience that their "multi-layered cognitive computing network" will be crucial in advancing AI towards GAI.
The supercomputer network is equipped with several advanced types of chips, including NVIDIA L40S, H200, GB200 Backwell systems, as well as AMD Instinct and Geona chips. The ultimate goal of AGI (Artificial General Intelligence) is to reduce the need for data, thereby decreasing the computational and energy resources required, instead of relying on a single powerful system.
It's Just Around the Corner!
Initially, however, these powerful supercomputer networks will be necessary to bridge the gap between current data-driven learning and eventually achieving human-level reasoning capabilities. This system will process deep neural networks that mimic human brain functions. Alongside these will be the familiar LLMs and multimodal systems that analyze human behaviors such as speech and movement.
This global network of supercomputers will aggregate the computational power essential for this transition towards AGI, according to SingularityNET's CEO. The network will be managed by an in-house software, which will oversee and protect the entire system. The first of these supercomputers is set to go online as soon as next month.
Expanded Explanation
The transition from narrow AI, which is focused on specific tasks like image recognition or language processing, to General Artificial Intelligence, which can perform any intellectual task a human can do, is one of the most ambitious goals in technology today. While current AI technologies require massive datasets and significant computational power, GAI aims to operate more efficiently, mimicking the cognitive flexibility of the human brain.
SingularityNET’s approach leverages a network of supercomputers to create a multi-layered cognitive computing environment. This is a step towards developing AI that not only processes information but also understands and reasons like a human. The inclusion of advanced hardware, such as the latest GPUs from NVIDIA and AMD, is key to handling the complex tasks required during this transition.
As this technology progresses, it will likely transform industries that rely on automation, decision-making, and cognitive tasks. The vision of having an AI that can understand context, reason, and learn autonomously could bring unprecedented advancements in fields ranging from healthcare to robotics and beyond.
The imminent deployment of the first supercomputers in this network marks a significant milestone in the journey towards realizing GAI, potentially revolutionizing how we interact with and utilize AI in our daily lives.