Now that Microsoft CEO Satya Nadella has appointed a new CEO to run Microsoft’s biggest businesses, he has a little more time on his hands for other adventures. Beyond focusing on Microsoft’s technical work, Nadella is now turning to the ancient art of blogging to discuss Microsoft’s year ahead and why he thinks everyone needs to move “beyond the arguments of [AI] slop vs sophistication.”

Nadella’s first blog entry in “sn scratchpad” is all about Microsoft and other AI companies still needing to get a bunch of stuff right with AI. Chief among them is creating a new concept for AI that evolves the “bicycles for the mind” concept that Steve Jobs used to describe computers as tools in the ’90s. “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,” says Nadella.

Nadella wants to move beyond the usual AI slop arguments, because Microsoft is betting on getting everyone hooked on AI agents instead of the Office and Windows software that has powered so many industries for decades.

This speaks to the tension with AI models right now and the fear from creatives about being edged out by AI models that are capable of copying the style of artists, designers, filmmakers, and more. We’ve been using PCs as tools for decades to create art, write code, and beyond, but Microsoft and others now want us to rely on AI agents as the new tools for creation instead — even if a lot of what is generated is slop.

Microsoft has a vision of everyone using Copilot with our voices to create content, search for information, and discover how to use things. The problem is that the vision doesn’t match reality right now, and barely any of what Copilot promises to do actually works.

Microsoft is betting on improved AI models to help Copilot and its own AI offerings, just as Meta warns that you can’t trust your eyes to tell you what’s real anymore. While Nadella has been part of the OpenAI, Google, and Anthropic AI model battle of 2025, he now argues that it’s how people choose to apply AI instead of individual model power that ultimately matters.

“We will evolve from models to systems when it comes to deploying AI for real world impact,” says Nadella. These systems will have to take into consideration the societal impact they have on people and the planet, he says. “The choices we make about where we apply our scarce energy, compute, and talent resources will matter. This is the socio-technical issue we need to build consensus around.”

Nadella’s first sn scratchpad blog entry is brief, but it’s all about 2026 being a “pivotal year for AI.” The same could be said for 2025, but Nadella thinks the industry now has a “clearer sense of where the tech is headed” and how it will shape its impact on the world. We’ll have to revisit whether the tech industry gets AI right this year once 2027 rolls around, but Nadella is now promising to deliver more of his personal “notes on advances in technology and real-world impact” in future blog posts throughout 2026.