Where is Research Headed Anyway?

Attending Usenix ATC and OSDI '25, two of the world's premier computer systems conferences, was an intense and enlightening experience. With three parallel sessions - two for ATC and one for OSDI - keeping tabs on everything was a real challenge, making for an exceptionally busy week.

The conference truly lived up to its reputation for rigorous academic focus, which is precisely as it should be. Academia's role is to push the boundaries, staying ahead of industry trends, and the papers presented here were definitely on the bleeding edge of computer systems research.

A striking observation was the overwhelming representation from universities in the US and China, together accounting for over 90% of participating institutions. China, particularly Tsinghua University, seemed to lead in the number of accepted papers. While there were some excellent contributions from Europe (primarily Switzerland) and South Korea, the presence from other scientifically inclined economies like India and Japan was notably sparse. This makes me wonder if researchers from these regions don't participate as much, or if significant work in computer systems isn't happening there. It's encouraging, though, that IIT Kanpur and IISc Bangalore had a couple of papers accepted in the ATC. China's deep focus on scientific research is something I'm keen to explore further, perhaps during a future visit.

The conference also provided invaluable networking opportunities. I had the privilege of meeting Atul Adya, an IIT Delhi B.Tech, MIT MS, and PhD alumnus. Our one-on-one conversation revealed he was advised by the legendary Barbara Liskov, and he shared insights into his recent work, Understanding the limitations of pubsub systems. As a distributed systems expert, he showed keen interest in our paper and also served as our session chair. It was also a pleasure to meet Murat Demirbas, whose blogs have taught me so much about distributed systems. Additionally, I connected with Zachary Yedidia, advised by D. Mazieres, whose significant paper in the OSDI day 2 session was truly impressive.

Our presentation, the first in the ATC track 1, was chaired by Atul Adya. We received numerous insightful questions during and after the session, and it was gratifying to see genuine interest in our work. I'm hopeful this will translate into future citations.

Even as someone specializing in computer systems, I found it incredibly challenging to fully grasp the breadth of work presented. Computer systems itself is a vast field, encompassing numerous niches like AI systems, compilers, hardware accelerators, consensus, verification, security, storage, operating systems, cloud computing, and blockchain. Each of these requires years, even decades, to master. It's practically impossible to delve deeply into every sub-field, and I often found myself struggling to follow presentations, even with my full attention.

This brings up a larger point. Even in a field like computer systems, which isn't considered "hot" compared to AI, the conference received over 1,000 papers. AI conferences like NeurIPS routinely see more than 3,000 submissions, making it seem easier to gain citations there. But this is just computer science. Consider the hundreds of engineering domains, health fields, and arts disciplines - millions of papers published annually across countless conferences. Does anyone truly have a handle on where the world of research is heading? A computer science PhD student at MIT might master a niche so specialized that no one else understands it without significant effort. With millions of such students, who has the time or energy to track it all? Perhaps this uncertainty is where the beauty of research lies. No one truly knows the future, but everyone is striving to make their unique mark, contributing to a collective reshaping of our world.

The era of groundbreaking, "relativity-level" papers seems to be behind us; current research largely focuses on incremental improvements. However, this incremental approach can sometimes obscure the larger end-to-end vision and the ultimate direction of research.

Beyond the academic discussions, the conference served as a prime recruiting ground. Leading computer systems research teams from Google, Meta, Databricks, Jane Street Tech, and Amazon Research were all present, eager to recruit the brightest PhD students who had made it to this prestigious event.

Overall, the experience was incredibly enriching. I learned a tremendous amount and am grateful to the reviewers and shepherd of our paper for the opportunity to contribute to the research community.

Looking forward to what's next!