SoftBank’s Son says AI is designing OpenAI’s next model
SoftBank CEO Masayoshi Son and OpenAI CEO Sam Altman attend an event to introduce artificial intelligence to businesses in Tokyo, Japan, on February 3, 2025.
Kim Kyung-Hoon | Reuters
The fact that OpenAI’s next model is being designed by another model is a sign that artificial intelligence has reached “superintelligence.” SoftBank CEO Masayoshi Son told CNBC.
The billionaire’s comments came suddenly warning AI development may need to be slowed down to deal with the consequences of the rapid pace of recovery, Anthropic said.
Son runs SoftBank, one of the world’s largest technology investors and one of the largest OpenAI shareholders. In an interview with CNBC on Monday, Son said he spoke with OpenAI CEO Sam Altman and engineers at the firm, who told him an AI model was “designing” a future model.
“This will happen to all other major models as well,” Son said, adding that engineers will no longer be smart enough to design the next model.
“So once that happens, [the] produces models [the] The next model… and it will be exponentially smarter than all of us. This is a superintelligence,” Son told CNBC.
An OpenAI spokesperson declined to comment on unreleased models but highlighted areas where the company is already using AI in model development.
In February, OpenAI said the GPT‑5.3‑Codex was its “first model that is effective at creating itself.” The team behind Codex, OpenAI’s coding tools, “used older versions to debug its own training, manage its own deployment, and diagnose test results and evaluations.”
‘Artificial super intelligence’
Son’s comments are part of a broader discussion about the term “artificial superintelligence,” or ASI, which he defines as artificial intelligence in 2024. 10,000 times smarter than humans. At the time, Son said ASI would be here in 10 years.
But he told CNBC on Monday that he was “trying to be conservative because people were shocked” when he laid out that timeline nearly two years ago.
“In my mind, I thought it would come in 4 years instead of 10 years. Now I say it will come in the next two years,” Son said.
The SoftBank CEO said he currently uses OpenAI’s ChatGPT two to three hours a day because the AI is smarter than him “on most things.”
Son said that within the next few years, AI will be smarter than humans in about 70% to 80% of subjects, and “could be 10 times smarter than average humans” on subjects where it surpasses human intelligence.
He’s been bullish on AI for the past few years, and his ownership of the chip designer puts SoftBank right in the middle of the boom Arm and shares in OpenAI, as well as investments in areas such as robotics and autonomous driving.
He told CNBC that the AI revolution is 50 times bigger than the dot-com revolution of the 2000s.
Anthropic warnings
The dangers of more advanced AI systems came to the fore Thursday after Anthropic published a blog post about “recursive self-improvement,” or RSI, a trend in which an AI system “has the ability to design and develop its own successor in a completely autonomous manner.”
While Anthropic said there would be positive results, it warned that “fully recursive personal development can also increase success.” risks humans losing control over AI systems.”
The company that developed the AI chatbot called Claude said a coordinated effort among AI labs to slow the development of this technology “would probably be a good thing.”
It is unclear whether Son was referring to RSI when talking about OpenAI’s model development. But an OpenAI study in June paper He said there are “early signs” of RSI in today’s systems.
“We expect this to increase competitive pressure between developers and countries and create governance challenges that existing institutions cannot address. As RSI emerges, societies will need ways to shape the course of AI development and ensure it serves human interests,” the article said.
