Microsoft falls as it lays out plan for “Humanist Superintelligence” without OpenAI

Microsoft stock is down 2% after the company announced it is developing its own superintelligence team to pursue AI that surpasses humans, but also keeps humans in charge. The so-called “Humanist Superintelligence” (not to be confused with Meta’s regular superintelligence) will have “incredibly advanced AI capabilities that always work for, in service of, people and humanity more generally,” AI Chief Executive Mustafa Suleyman wrote.

It’s also Microsoft’s first attempt at achieving artificial general intelligence without OpenAI, which it’s now free to do after it renegotiated their partnership. In an interview with The Wall Street Journal, Suleyman praised the companies’ work together but also criticized chatbots that feel too human and trick people into thinking they’re sentient — a criticism others have leveled at OpenAI’s ChatGPT, which is widely considered to be leading the way with AI. While the two companies are continuing to work together, it’s also clear that Microsoft thinks it can make AI progress on its own.

AI is also very expensive — something that’s given investors pause lately as they wonder just how companies will pay for their giant capital outlays and how those will ultimately pay off. Meta last week took a big dive, partly because investors don’t see a clear route from the social media company’s massive AI infrastructure spending to revenue, since it’s not renting that infrastructure as a cloud business.

Even Microsoft, which has a robust cloud business, hasn’t escaped investor scrutiny over AI spending. Despite beating revenue and earnings expectations last week, the company’s stock fell in part because the company reversed its guidance on capital spending, now saying it would continue to rise instead of taper off.

Regardless of how humane it might be, the pursuit of superintelligence — and the talent and infrastructure it involves — will cost those who dare try it. And the payoffs are still unclear.