According to CNBC, Microsoft announced on Thursday that it’s forming a new MAI Superintelligence Team under AI chief Mustafa Suleyman, who will personally lead the advanced artificial intelligence research effort. The decision comes just months after Facebook parent Meta spent billions hiring talent for its own Meta Superintelligence Labs unit. Suleyman, a DeepMind co-founder who later started Inflection before Microsoft hired him and several colleagues last year, wrote in a blog post that the team aims to solve “real concrete problems” while remaining “grounded and controllable.” Microsoft, which uses OpenAI models in Bing and Copilot while owning a $135 billion equity stake in OpenAI, has been taking steps to reduce its dependence on the ChatGPT maker by also using models from Google and Anthropic.
The superintelligence arms race is here
So we’re officially in the superintelligence era now. Microsoft‘s move basically confirms what many in the industry have been whispering about – that the big tech players are all scrambling to be first to achieve AI that surpasses human intelligence. Meta’s already spending billions, and now Microsoft is throwing its hat in the ring with a dedicated team.
Here’s the thing though – this isn’t some academic research project. Suleyman specifically mentioned they’re focusing on “narrow areas in medicine and renewable energy production” and want “expert level performance at the full range of diagnostics.” That sounds incredibly ambitious. But it also raises the question: what happens when you’re building systems that could literally outperform the world’s best doctors and energy experts?
Suleyman’s interesting track record
Now let’s talk about the man leading this charge. Mustafa Suleyman has quite the resume – co-founded DeepMind, sold it to Google, left to start Inflection, then Microsoft hired him and essentially absorbed much of his startup. That’s a pattern worth noting. Microsoft isn’t just building this capability from scratch – they’re acquiring talent and expertise through strategic hires.
And honestly, Suleyman’s background gives him some credibility here. DeepMind was doing cutting-edge AI research long before ChatGPT made it mainstream. But his departure from Google in 2022 and the quick pivot to Microsoft raises questions about how these superintelligence projects will actually play out in the corporate environment.
Practical technology or perilous pursuit?
Suleyman’s blog post is filled with reassuring language about building “practical technology explicitly designed only to serve humanity” and not pursuing superintelligence “at any cost, with no limits.” That sounds great in a press release, but the reality is we’re talking about systems that could fundamentally reshape society.
I mean, think about it. We’re still struggling with basic AI safety issues – hallucinations, bias, misinformation – and now we’re racing toward superintelligence? The gap between “we promise to be careful” and actually having robust safety measures feels pretty wide right now.
And let’s not forget the business pressures. Microsoft investors are already getting nervous about AI spending without clear profit paths. Suleyman had to address that directly in his announcement. But when you’re in a race against Meta and others, how long before the “at any cost” mentality creeps in?
What comes next?
Microsoft’s approach seems more measured than some competitors – focusing on specific domains like medicine and energy rather than general superintelligence. That’s probably smart. But the term “superintelligence” itself sets incredibly high expectations.
The company’s official announcement talks about creating “useful companions for people” in education and other areas. That sounds benign enough. But we’ve seen how quickly AI capabilities can surprise even their creators.
So while Microsoft promises a grounded, controllable approach, the history of technology suggests we should maintain some healthy skepticism. Superintelligence isn’t just another product feature – it’s potentially civilization-changing technology. Let’s hope the reality matches the reassuring rhetoric.
