Ratul Mahajan

I SEE (study, engineer, educate) computer networks. Tweets are facts; retweets are alternative facts

[Ratul Mahajan @ratulm]

@BenjDicken Thanks for the shoutout. Giacomo is here as @mino98; he prefers to just lurk. (Sorry that I outed you, Giacomo.)

[Ben Dicken @BenjDicken]

Combining deep math theory with practical implementation leads to many cool things.

This is the heart-and-soul of computer science and engineering.

AWSs latest blog is a beautiful example, describing and visualizing the transformation from tree-based networks to pseudo-random graphs in their DCs.

Props to @ratulm and the other authors (who need to get on X dot com ASAP)

[Matt Garman @mattsgarman]

Most people thought flat data center networks would never work at hyperscale.

The @awscloud team figured it out.

Resilient Network Graphs are a completely new network architecture, now live.

33% better throughput, 40% less network power.

https://t.co/LEaIItVqe9

[Amazon Science @AmazonScience]

A passive optical device called a ShuffleBox and a routing protocol called Spraypoint just made flat datacenter networks as practical as fat trees, with 69% fewer routers and up to 33% better throughput. https://t.co/M9mEzCWfrx

[Ratul Mahajan @ratulm]

The real story: https://t.co/DTx2ae3sl4

[Ratul Mahajan @ratulm]

Look, Ma, no more trees!

https://t.co/F4Wy59gFB9

[Ed Lazowska @lazowska]

Huge congratulations to #UWAllen's Magda Balazinska and Shwetak Patel! / Three UW faculty members elected to the American Academy of Arts and Sciences https://t.co/X072KHbO00

[Ratul Mahajan @ratulm]

Thought provoking from @om as usual.

Your digital twin "is a monument to a self that stopped growing."

https://t.co/mLxgGQIcSi

[Ratul Mahajan @ratulm]

Is it me or is this as stupid as it sounds?
https://t.co/5rKUpmo3OT

[Ratul Mahajan @ratulm]

Grateful to @LaudeInstitute for the honorable mention award for our self-defining systems project, which focuses on building and operating large computer systems using AI.

Read about our vision at https://t.co/CbiDLAnxoR (w/ Tom Anderson, Simon Peter, @LukeZettlemoyer)

[Laude Institute @LaudeInstitute]

We asked the best AI researchers in the world one question: if you had the resources, how would you use AI to solve humanity's hardest problems?

Today we're sharing what came back.

Moonshots // ONE is live. 🧵 https://t.co/tMRWDNOXaU

[Ratul Mahajan @ratulm]

UW systems seminar this quarter is focusing on building reliable systems using LLMs.

If you want to follow along, see https://t.co/H7Z7f9ffHe

[Matei Zaharia @matei_zaharia]

Registration is open for the CAIS conference! We have some awesome workshops and papers in the program.

[Ratul Mahajan @ratulm]

If leaders follow their dreams, what does that make them—followers?

[Ratul Mahajan @ratulm]

@ShriramKMurthi Now do the UK

[Ratul Mahajan @ratulm]

me: /context

claude code: 54k/1000k tokens (45%)

me: how is that right?

cc: It's not — looks like a rendering bug.

[Vic Shihang Li @sudopowr]

Today's AI agents can diagnose production incidents, but they start from scratch every single time. What if they could remember?
New on @acmsigops: our work on the Self-Defining Operator, a multi-agent system with long-term memory for autonomous ops.

[Tianyin Xu @tianyin_xu]

Very interesting work on Self-Defining Operator which develops long-term memory to resolve operational issues faster. SDO is one instance of Self-Defining Systems (SDS) led by Vic, @ratulm and colleagues at @uwcse. Love the system-centric vision!

[Ratul Mahajan @ratulm]

Early report from the first project we are pursuing as part of the "self-defining systems" research, led by @sudopowr.

So much potential to lower operational toil and MTTR!

[ACM SIGOPS @ACMSIGOPS]

New SIGOPS Blog -- "The Long Game: How Agents That Remember Resolve Operational Issues Faster" by Shihang (Vic) Li, Thomas Anderson, Ratul Mahajan, Simon Peter, Luke Zettlemoyer, and the SDS team.
https://t.co/mjYKGKWRf9

[Ratul Mahajan @ratulm]

A beautiful, haunting essay by @om. Might make you uneasy with yourself and your wants.

https://t.co/t9l7ijXrCd

[Ratul Mahajan @ratulm]

AI coding is a great equalizer for superficial code.

But for complex systems, it widens the gap — between those who design before they build, and those who discover through iteration.