Steadfast, at least for the moment, in Microsoft’s headlong descent into AI and generative AI, CEO Satya Nadella has invoked the ancient ways of blogging in his latest defense of the technology and its unchecked, profit-resilient deployment, insisting with ‘too big to fail’ energy that we must move on from the slop fiasco because of the reasons why.
In a December 29, 2025 post to his “sn scratchpad” blog, Nadella argues 2026 “will be a pivotal year for AI. Yes, another one. But this moment feels different in a few notable ways.” I agree; this time there’s a blog.
“We now have a clearer sense of where the tech is headed, but also the harder and more important question of how to shape its impact on the world,” he continues, not pausing to specify or, really, consider those impacts or how to shape them, be they environmental, labor, or economical consequences that reckless AI investment has already yielded. Ownership, misinformation, political weaponization – there are myriad impacts to put on the scales opposite the ability to futilely ask a computer to make you feel something or do your job.
With the tremendous weariness of a man surely tired of watching AI spectacularly lose arguments of slop, Nadella insists “we need to get beyond the arguments of slop vs sophistication and develop a new equilibrium in terms of our ‘theory of the mind’ that accounts for humans being equipped with these new cognitive amplifier tools as we relate to each other.” On this I firmly agree; I, too, seek to avoid losing battles.
This echoes the stance of Microsoft AI CEO Mustafa Suleyman, of course, who said in November that he simply cannot believe some people are underwhelmed, nay, unimpressed with the AI wonders they’ve cooked up, as ever conflating pattern recognition with “conversation” and hasty, unavoidably derivative slop for innovation. It’s so often the incredulity, the spoken or implied ‘my way or the highway’ absolute, that truly shakes these claims.
“We will evolve from models to systems when it comes to deploying AI for real world impact,” Nadella declares for 2026, tacitly saying that “we” have not done that. To that end, “We have learned a lot in terms of how to both keep riding the exponentials of model capabilities, while also accounting for their ‘jagged’ edges,” Nadella says, not specifying what those jagged edges are or how sharp they might be.
At last, Nadella approaches a salient point: “We need to make deliberate choices on how we diffuse this technology in the world as a solution to the challenges of people and planet.” Deliberate is not how I’d describe the diffusion of AI thus far, not from Microsoft or many others. Helter-skelter, perhaps, even desperate. Preachy. Expectant. Yet “for AI to have societal permission it must have real world eval impact,” Nadella says, bringing us back to the sticking point of slop and jagged edges versus real, positive impact. Shoutout to the architects of the experimental AI pattern recognition systems working to identify cancers early, for instance.
“It will be a messy process of discovery, like all technology and product development always is,” Nadella adds, almost hedging, or at least hiding behind the shield of history to imply catastrophic growing pains should be expected and accepted no matter the cost. The ends justify the means. Not yet, obviously, but surely they will this year.
“Computing throughout its history has been about empowering people and organizations to achieve more, and AI must follow the same path,” he concludes. “If we do that, it can become one of the most profound waves of computing yet. This is what I hope we will collectively push for in ‘26 and beyond.”
AI certainly could become something profound – it has already materially impacted games under Microsoft’s umbrella – but as always, I’m stuck on the gulf between what’s described and what’s delivered. Perhaps the “discovery” phase Nadella is eager to move on from has been a bit too short. Microsoft has essentially discovered the AI equivalent of the Model T automobile, yet speaks as if it has already shipped the McLaren W1, and hand-waves any concerns about steps that may have been skipped.
For the creators of Baby Steps, one of 2025’s best and weirdest games, there’s no room or reason to surrender creativity to AI: “The player is in communion with a human designer.”
