AI's role in operating systems is a controversial topic, and one that sparks passionate debates. But is there a place for AI within an OS? The answer, according least to some, is a resounding no.
A recent online post by Windows leader Pavan Davuluri, celebrating Windows' transformation into an 'agentic OS', faced a backlash in the comments section. Users demanded reliability, usability, and stability, seemingly dismissing the idea of an agentic OS. Davuluri's response acknowledged the feedback but hinted at a hidden agenda, suggesting that user feedback isn't the only factor in play.
Engineering, the art of compromise and innovation, has its rules. Prioritizing fundamentals is essential, but Microsoft's focus seems to be shifting. An agentic OS, a vague concept, appears to be the priority, potentially compromising the core principles of an operating system.
An OS has a defined role: controlling computer resources and presenting them as standard services to applications and users. From MS-DOS 1.0 to Windows 11, the OS has evolved, adapting to new architectures, multitasking, and security. Yet, its primary function remains unchanged.
Agentic computing, however, doesn't align with this. It's not an OS service and may even contradict OS principles. It's about agents performing tasks for users across local and remote services. An agentic platform makes sense, but an 'agentic OS' is a confusing concept, especially in desktop computing.
The placement of agentic functionality within the OS raises concerns. It implies privileged access to data and services, which is unnecessary for user tasks. Microsoft's tendency to blur the lines between the Windows OS and its bundled apps and services further complicates matters.
The insistence on integrating user-level applications into the OS core, as seen in the Internet Explorer antitrust case, raises eyebrows. This strategy, driven by corporate ambitions, is not new. But the current AI dominance scramble demands critical thinking.
The push for agentic AI seems to originate from sources other than user feedback. While users must use an OS, they can choose to adopt a platform. An agentic platform, however, may struggle to gain traction.
In summary, AI's role in operating systems is questionable. While reliability, usability, and stability are essential, the concept of an agentic OS seems to be a controversial and potentially misguided endeavor. But is this the whole story? Share your thoughts in the comments below!
