AI is designing OpenAI’s next model in a sign of ‘superintelligence’: SoftBank’s Masayoshi Son to CNBC SoftBank CEO Masayoshi Son revealed that an AI model is being used to design OpenAI’s future models, a significant step towards what he calls ‘superintelligence.’ He believes this phenomenon, where AI creates its successor, will lead to AI becoming exponentially smarter than humans. This rapid advancement, also discussed by Anthropic regarding ‘recursive self-improvement,’ prompts concerns about human control and a potential need to slow down AI development.

• An AI model is reportedly designing OpenAI’s next generation of models.
• SoftBank CEO Masayoshi Son believes this indicates AI is approaching ‘superintelligence’.
• Son has revised his timeline for superintelligence, now expecting it within two years.
• Anthropic has warned about the risks of ‘recursive self-improvement’ (RSI) in AI, suggesting a slowdown in development.
• OpenAI previously stated that GPT-3.5 Codex was instrumental in its own creation.
• Son views the AI revolution as significantly larger than the dot-com boom.