{"id":25210,"date":"2025-07-09T17:30:00","date_gmt":"2025-07-09T14:30:00","guid":{"rendered":"https:\/\/forklog.com\/en\/a-neural-network-of-digital-mind-copies\/"},"modified":"2025-07-09T17:30:00","modified_gmt":"2025-07-09T14:30:00","slug":"a-neural-network-of-digital-mind-copies","status":"publish","type":"post","link":"https:\/\/forklog.com\/en\/a-neural-network-of-digital-mind-copies\/","title":{"rendered":"A neural network of digital mind copies"},"content":{"rendered":"

In late June, OpenAI co-founder Ilya Sutskever shared<\/a> his vision for artificial intelligence. He believes AI agents will soon complete a month\u2019s worth of human work in hours and become fully fledged virtual colleagues, scalable to millions. The shift to models with genuine \u201creasoning\u201d that \u201cchange everything\u201d unnerves many. Sutskever urged people not to fear the future but to shape it by adapting to change.<\/p>\n

In this ForkLog piece, we examine how cognitive AI systems are evolving\u2014and the challenges they pose not only to developers but also to users.<\/p>\n

Think about thinking<\/h2>\n

Over the past two years the AI community has increasingly debated the consequences of parabolic progress, the arrival of artificial general intelligence (AGI) and even superintelligent ASI<\/span>. In that context, researchers are trying to define the next step.<\/p>\n

Most corporate \u201cseers\u201d predict that the next generation of AI will take the form of various flavours of cognitive artificial intelligence (CAI).<\/p>\n

Google DeepMind chief executive Demis Hassabis told Wired there is<\/a> a 50% chance of achieving AGI within 5\u201310 years. Beyond that, he said, lies an \u201cera of maximum human flourishing\u201d, with the colonisation of the galaxy starting as early as 2030.<\/p>\n

Google\u2019s Gemini chatbot\u2019s Deep Think mode may be the next step in \u201cmetacognition\u201d, in which a machine performs multi-layer logic and planning, edging closer to cognitive thought.<\/p>\n

Meta AI\u2019s chief scientist Yann LeCun a year ago urged the next generation to refrain from studying<\/a> LLM<\/span>s. He argues that big companies already dominate the field and that it is not the path to advancing AI.<\/p>\n

\n

\u201cIf you\u2019re a student who is interested in building the next generation AI systems, don\u2019t work on LLMs,\u201d<\/em> he said.<\/p>\n<\/blockquote>\n

Instead, Meta\u2019s team has focused on \u201cworld models\u201d like those in animals and humans\u2014the V-JEPA<\/a> system, designed to form a genuine understanding of the environment. It learns intuitive reasoning from video using self-supervised learning. Compared with multimodal LLMs, it has shown a deeper grasp of physics while requiring less training data. The model makes predictions in an abstract representation space\u2014akin, in LeCun\u2019s view, to how the human brain processes information.<\/p>\n

Since early 2024, OpenAI has begun introducing long-term memory into ChatGPT, a step toward modelling a persistent \u201cself\u201d and an individual thinking style. The firm ultimately aims to build a network of AI agents with deep user understanding.<\/p>\n

One of the main infrastructure suppliers for global AI\u2014Nvidia\u2014has launched a set of initiatives to advance robot spatial orientation technology\u2014the DePAI<\/a> stack. Chief executive Jensen Huang regards modelling the real world and its digital replicas, including people, as a precursor to reasoning AI.<\/p>\n

In short, CAI is an attempt to embed in systems the following qualities:<\/p>\n