AI inauguration SingularityNet is set to deploy “ multi - level cognitive computing web ” in the coming months . It is designed to host and train the models that will form the base of anartificial general intelligence(AGI ) capable of matching — even potentially exceeding — human knowledge , the company announced on Monday .
accomplish AGI is wide viewed as the next major milepost in artificial intelligence activity developing . While today ’s cut - edge models like GPT-4o and Gemini 1.5 Pro are immensely hefty and can perform specific tasks at superhuman levels , they ’re incapable of applying those attainment across discipline . AGI , though still theoretic at this point , would be spare of those limitations , and capable to conclude and learn on its own , irrespective of the task .
SingularityNet is go to build up the compute base necessary to condition and deploy such a organization using some of the most advanced components currently on the grocery . Per a report fromLiveScience , the inauguration ’s modular supercomputer will sport Nvidia L40S GPUs , AMD Instinct and Genoa processors , Tenstorrent Wormhole server rack unravel Nvidia H200 GPUs , as well asNvidia ’s 1,500W - plus GB200 Blackwellsystems .
Supercomputer architectures differ from your formal screen background setup in that they range multiple sets of processor ( both CPU and GPUs ) assembled into individual thickening . Those nodes then get daisy - chained together by the tens of thousands into the larger arrays of the overarching supercomputer .
“ This supercomputer in itself will be a breakthrough in the transition to AGI . While the novel neural - symbolic AI approaches recrudesce by the SingularityNET AI team lessen the need for data , processing and push somewhat relative to standard bass neural nets , we still need significant supercomputing facilities , ” SingularityNet CEO Ben Goertzel told LiveScience . “ The mission of the computer science machine we are creating is to ensure a phase transition from learning on grown datum and subsequent reproduction of setting from the semantic memory of the neural electronic connection to non - imitative machine cerebration based on multi - step logical thinking algorithms and active humankind modeling . ”
“ Before our center , a paradigmatic shift is taking position towards continuous learning , unlined generalisation and reflexive AI ego - modification , ” he proceed .
The troupe be after to Grant public access to the supercomputer , once it comes fully online in belated 2024 / early 2025 , using a nominal system . Users will buy tokens , as they would in an honest-to-god - schoolhouse colonnade , and then pass those tokens to get a certain turn of opportunity to play with the system . The data generated by those interactions is then feed back into the organization for further AGI experiment and development .
SingularityNet is far from the only company rush to build and deploy the first AGI . The pursuit of such systemsis one of OpenAI ’s founding tenetswhile Meta ’s Mark Zuckerberghas appropriate more than $ 10 billionfor his companionship ’s AGI R&D.
