GabyThis forecast was announced by leading AI scientist Ben Goertzel at the Beneficial AGI 2024 summit in Panama.
BenHerzl– dDoctor of Philosophy, mathematician, futurist, and CEO.SingularityNET, whichin the early 2000sintroduced the term AGI: “artificial general intelligence.”Since then, he has been called the “father of AGI.”
Therefore, according to the scientist, research in the field of AI We are entering a period of exponential growth. New data suggests that AGI, where artificial intelligence becomes as capable as humans in several areas, regardless of the training data, is already within reach. This hypothetical point in AI development is known as the “singularity.”
According to the futurist, singularity could occur in 2029-2030. However, it could happen even earlier: in 2027.
Ben Goertzel is a Doctor of Philosophy.
What c forecast It seems that the text you provided is incomplete or unclear. Could you please provide more context or a complete sentence for translation? Ben Herzl
The best systems artificial intelligence The systems used today are considered “narrow artificial intelligence.” While they may be more “advanced” than humans in a specific area, according to training data, they cannot surpass human intelligence overall. These AI systems, which range from machine learning algorithms to large language models (LLMs) like ChatGPT, are incapable of reasoning and understanding context like humans do.
Ben Goertzel suggested that humanity could likely create the first version of AGI around 2029-2030. However, it could happen as early as 2027. If this version of AI is designed to have access to its own code and rewrite it, it could quickly transform into artificial superintelligence. Goertzel generally defined it as AI that possesses the cognitive and computational power of all human civilization.
According to the scientist, no one has yet created human-level AGI; no one has a clear idea of when we will reach this. “On the other hand, it is quite possible that we could achieve human-level general artificial intelligence, say, in the next three to eight years,” Hertzell said.
One step away from singularities
To support his thesis, he pointed to “three lines of similar evidence,” the publication reported. Live Science The first is the modeling of computer scientist Ray Kurzweil in his books “The Singularity Is Near” (2005) and “The Singularity Is Closer” (2024). In his works, Kurzweil suggested that superintelligence will emerge as early as 2029.
Ben Goertzel also pointed out the improvements made to LLM programs over the years, which have opened up the potential of AI to the world. The expert explained that LLMs alone will not lead to AGI. After all, the way they demonstrate knowledge does not reflect true understanding. However, LLMs can be one of the components of a broad set of interconnected architectures.
The third piece of evidence, according to Hertzel, is his work on creating the infrastructure he called OpenCog Hyperon, as well as the related software systems and the future programming language for AGI, named MeTTa.
OpenCog Hyperon is a form of AI infrastructure that aims to integrate both established and new paradigms of artificial intelligence. The hypothetical endpoint is a large-scale network of AI systems based on various architectures, each representing different elements of human cognition: from content generation to reasoning.
Hertzell, however, acknowledged that he “could be wrong” and that humanity might need “a quantum computer with a million qubits or something like that.” “Once you reach human-level AGI, within a few years you could achieve radically superhuman AGI – unless it stifles its own development due to its own conservatism,” Mr. Hertzell added.
