Masayoshi Son is known for taking big bets, like those with Alibaba and WeWork, and he now seems to have taken another big bet — against human coding.
Softbank CEO Masayoshi Son has said that he believes that at least within Softbank, the era of human coding is coming to an end. “The era when humans program is nearing its end within our group. Our aim is to have AI agents completely take over coding and programming,” he said at a customer event. Son added that the Softbank group had already initiated the process for implementing his vision.
Son said that Softbank estimated that it needed to create around 1000 AI agents per person – a large number because “employees have complex thought processes. “The agents will be active 24 hours a day, 365 days a year and will interact with each other,” he said.
Son said that these AI agents will operate at a fraction of the costs of the humans they’d be replacing. He estimated that each agent would cost just 40 yen ($0.27) per month. With a thousand agents replacing a programmer, it would work out to just $270 per month. This is significantly less than what human coders typically get paid. Son said that his agent plan would save the company $3.2 billion per year.
“For 40 yen per agent per month, the agent will independently memorize, negotiate and conduct learning. So with these actions being taken, it’s incredibly cheap,” Son said.
Son didn’t provide many details on when or exactly how his vision of these agents replacing human programmers would be a reality, but he admitted that it would be “challenging” to achieve. He admitted that Softbank hadn’t yet developed the necessary software to support having millions of AI agents.
Son claimed the capabilities of AI agents had already surpassed PhD-holders in advanced fields including physics, mathematics and chemistry. “There are no questions it can’t comprehend. We’re almost at a stage where there are hardly any limitations,” he said. He also dismissed the problem of AI hallucinations, saying that they were a “temporary and minor issue”.
These seem like bold pronouncements, but Masayoshi Son has also put his money where his mouth is. Softbank has announced that it was leading a $500 billion investment into OpenAI’s Stargate datacenter. It was a mind-boggling amount, and had led Elon Musk to remark that it was a bluff, and Masa didn’t actually have the money. When Masa had similarly announced a $1 trillion plan for a robotics industrial complex, Musk had said that his “bullshit was reaching epic proportions”.
Masa Son, though, does seem to genuinely believe in the potential of AI. He’s spoken about how it could be worth spending “trillions” on AI because even if it ends up accounting for 5 percent of the global GDP, it could be a sound investment. He also says that AI can become “billions” of times smarter in a few years, and that AGI can be reached much earlier than 2027. Now all of Masa’s pronouncements haven’t quite worked out — he’s had some big duds in his career, including WeWork — but with him saying that Softbank is already laying the ground to replace coders with AI agents, it does seem that he’s serious about how AI could transform work and jobs in the coming years.