Home
Masayoshi Son: Superintelligence to Hit Critical Point in Two Years as AI Self-Iteration Loop Closes

Recently, SoftBank Group CEO Masayoshi Son made a startling prediction that has reverberated across the industry: the arrival of super artificial intelligence (ASI) will be greatly accelerated, with its official debut expected within just two years.
Notably, this marks the second time this month Son has revised his timeline for the technology's breakthrough. He has adjusted his predictions multiple times before—first estimating ten years, then four years, and now confidently placing this critical milestone within two years.
This bold prediction is rooted in a fundamental shift in AI's underlying R&D paradigm. Son revealed that after in-depth discussions with the OpenAI team, he learned that the self-iteration closed loop in AI has been initially completed. This means current large AI models are no longer entirely reliant on human programmers; they can now independently design and train the next generation of models. With this "recursive self-iteration" underway, AI intelligence will no longer grow linearly but will instead experience an alarming exponential surge, ultimately surpassing human research capabilities entirely.
Emerging data from the industry strongly supports this view. Several top global tech institutions have already identified related technical prototypes in their labs. Data released by competitor Anthropic shows that over 80% of the code on its platform is now written entirely by Claude AI autonomously. AI has evolved from a supporting tool into a core R&D force. Meanwhile, OpenAI has confirmed that its GPT-5.3-Codex model has achieved meaningful self-iteration, enabling it to independently handle the entire R&D workflow—from training diagnostics and deployment monitoring to final outcome evaluation.
However, this technological singularity—driven by AI itself—has sparked widespread concern among the tech community and global regulators. In its latest academic paper, OpenAI issued a public warning that the current rapid evolution of AI, particularly its "self-reproduction" capability, has far outpaced human society's expectations. The existing global social systems, laws, and regulatory frameworks are not yet ready for the arrival of super intelligence. As the self-iteration loop of AI closes, building a comprehensive control and guidance mechanism to prevent technology from spiraling into uncontrollable territory has become an urgent common challenge for the global tech community.
Related article
DeepMind CEO Hassabis: I sleep six hours a day, usually feel energetic around 1 a.m.
Fortune recently featured an interview with Demis Hassabis, CEO of Google DeepMind, revealing his unconventional approach to rest and productivity. Hassabis disclosed that he sleeps very little, structuring his waking hours into two distinct work blo
OpenAI, Anthropic Vie for Market Share Despite Revenue Shortfalls
Despite recent reports suggesting OpenAI missed revenue targets, creating pressure on tech stocks this Tuesday, private AI lab investors remain resilient. Seasoned backers have confirmed they will not reduce investment despite negative media coverage
California AV Compliance: A New Era of Tickets, Geofences, and 1M Miles
Guident operates an AuveTech shuttle in South Florida, managing a four-mile route in West Palm Beach and a one-mile route in Boca Raton using its remote monitoring technology. | Credit: GuidentCalifornia is redefining the regulatory landscape for dri
Related Special Topic Recommendations
Comments (0)
0/500

Recently, SoftBank Group CEO Masayoshi Son made a startling prediction that has reverberated across the industry: the arrival of super artificial intelligence (ASI) will be greatly accelerated, with its official debut expected within just two years.
Notably, this marks the second time this month Son has revised his timeline for the technology's breakthrough. He has adjusted his predictions multiple times before—first estimating ten years, then four years, and now confidently placing this critical milestone within two years.
This bold prediction is rooted in a fundamental shift in AI's underlying R&D paradigm. Son revealed that after in-depth discussions with the OpenAI team, he learned that the self-iteration closed loop in AI has been initially completed. This means current large AI models are no longer entirely reliant on human programmers; they can now independently design and train the next generation of models. With this "recursive self-iteration" underway, AI intelligence will no longer grow linearly but will instead experience an alarming exponential surge, ultimately surpassing human research capabilities entirely.
Emerging data from the industry strongly supports this view. Several top global tech institutions have already identified related technical prototypes in their labs. Data released by competitor Anthropic shows that over 80% of the code on its platform is now written entirely by Claude AI autonomously. AI has evolved from a supporting tool into a core R&D force. Meanwhile, OpenAI has confirmed that its GPT-5.3-Codex model has achieved meaningful self-iteration, enabling it to independently handle the entire R&D workflow—from training diagnostics and deployment monitoring to final outcome evaluation.
However, this technological singularity—driven by AI itself—has sparked widespread concern among the tech community and global regulators. In its latest academic paper, OpenAI issued a public warning that the current rapid evolution of AI, particularly its "self-reproduction" capability, has far outpaced human society's expectations. The existing global social systems, laws, and regulatory frameworks are not yet ready for the arrival of super intelligence. As the self-iteration loop of AI closes, building a comprehensive control and guidance mechanism to prevent technology from spiraling into uncontrollable territory has become an urgent common challenge for the global tech community.
DeepMind CEO Hassabis: I sleep six hours a day, usually feel energetic around 1 a.m.
Fortune recently featured an interview with Demis Hassabis, CEO of Google DeepMind, revealing his unconventional approach to rest and productivity. Hassabis disclosed that he sleeps very little, structuring his waking hours into two distinct work blo
OpenAI, Anthropic Vie for Market Share Despite Revenue Shortfalls
Despite recent reports suggesting OpenAI missed revenue targets, creating pressure on tech stocks this Tuesday, private AI lab investors remain resilient. Seasoned backers have confirmed they will not reduce investment despite negative media coverage











