About

About Architectural Prisms

Architectural Prisms is an experimental initiative founded and led by Karu Sankaralingam. Its an experiment and is a bit like a provocative piece of art

Karu is the Mark D. Hill and David A. Wood Professor of Computer Sciences at the University of Wisconsin-Madison, where he leads the Vertical Research Group. His research focuses on computer architecture, microarchitecture, and the broad interplay between hardware and software. In addition to his academic work, he was the founder and CEO of SimpleMachines, Inc., an AI hardware company that developed novel chip designs for high-performance AI and data processing. Karu is also currently a Principal Research Scientist at NVIDIA. He is a Fellow of the IEEE and a recipient of numerous awards, including the NSF CAREER award and the IEEE TCCA Young Computer Architecture Award. This project stems from his deep involvement in both academic research and the practical, at-scale application of cutting-edge AI. This project is a personal project.

The Vision: Re-engineering Peer Review

This project was directly inspired by Karu's ideas on the future of academic critique, first expressed in his SIGARCH blog post, "The Reviewer is Dead, Long Live the Review: Re-engineering Peer Review for the Age of AI".

We believe the traditional, human-only peer-review system is facing unsustainable pressures. The rising volume of paper submissions, coupled with chronic reviewer fatigue, a frequent lack of deep expertise, and misaligned incentives, often struggles to produce the high-quality, objective feedback that research deserves.

Architectural Prisms is founded on the belief that Large Language Models (LLMs) offer a transformative solution. We believe modern LLMs have reached a level of technical understanding that can meet, and in many cases surpass, the consistency of human-only review. AI models are not mentally fatigued, nor do they possess the personal egos or emotional biases that can lead to unfair critiques.

This platform is our experiment to explore that potential. We are at the beginning of a major shift, and we believe that within the next 2-3 years, the scientific community will rapidly transition toward an AI-first review process. Our goal is to help build and shape that future. Join, chime in, share your opinions and contribute.