Lav Varshney

Excited to share our new paper Iterative Decoding of Stabilizer Codes under Radiation-Induced Correlated Noise is on arXiv:arxiv.org/abs/2603.18231. Thanks to my advisor @lrvarshney and thanks to Paul Baity and team at @BrookhavenLab.

Applications are now live! Cohort 0 starts March 13th in Presidio with OpenHands, OpenRouter, alphaXiv, Fireworks, Dedalus Labs, Franklin Templeton, Founders Fund and Pantera. → $25K+ in prizes → 3 weeks building state-of-the-art AI agents → Many more surprises Apply below 👇

In case one is curious, the idea of modal controllability was first developed by my late PhD advisor Sanjoy Mitter in 1968.

Methodologically, it was important that we consider both structural and functional connectivity, both anatomy and physiology.

New paper, "The network architecture of general intelligence in the human connectome" is out in Nature Communications today (link below), providing key insight into structural and functional aspects of the brain that are associated with individual differences in performance on intelligence tests. Our team did amazing work: Ramsey Wilcox, @BabakHemmatian , and @DecisionNeurosc! The general intelligence factor g is one of the most robust findings in psychology and is predictive of numerous social outcomes such as educational attainment, job performance, and health. It is traditionally thought of as a psychometric statistical abstraction, but we show it arises from structure and coordinated activity across the brain’s global network architecture.

New approaches to AI and its implementation can provide novel capabilities but improve water, energy, data, and heat efficiency.

I spoke to @nbcchicago about new data centers, their uses, and their potential impacts. I believe balancing costs and benefits is very much a local issue for each community to figure out nbcchicago.com/video/news/loc…

Federation, pooling, ensembling, complex constructions from individual models, etc.: lots of technical ways emerging for institutions to work together on AI!

This works even with extreme heterogeneity, e.g. where CLIP embeddings cluster by modality.

With our friends from University at Buffalo @lokhande_tweets and Lawrence Livermore National Laboratory, new paper led by Ayush Roy "Beyond Pooling: Matching for Robust Generalization under Data Heterogeneity" will be at AISTATS 2026 in Tangier, Morocco.

AI Model Helps Doctors Plan Safer Radiation Treatments, work presented at @MICCAI_Society 2025 Developed by , Zhaozheng Yin, student Jignesh Chowdary G., Dr. Tiezhi Zhang, and Dr. Xin Qian ai.stonybrook.edu/about-us/News/… @SUNY @stonybrooku @StonyBrookMed @sbucompsc @CEASSBU

I had recently spoken with Lin-Manuel Miranda. He was mentioning that his best ideas (like Hamilton) came from reading a thick book while sitting on a beach. I have a trip coming up and am looking for suggestions that can "Lin" up my creativity: please reply below. It was also very cool to get his take on AI creativity and the creator-focused, authentic way we do it at Kocree.

The @ENERGY has been charged with leading the "Genesis Mission" for advancing AI. Scientists from national laboratories, American businesses, and world-renowned universities are being asked to achieve dramatic acceleration in AI development and utilization. Executive Order: whitehouse.gov/presidential-a…

In an interview, Director @lrvarshney shared his initial experiences at @stonybrooku, and his vision and mission for the AI Innovation Institute, highlighting its role as a catalyst for discovery, dialogue, and action in AI in New York State, and beyond ai.stonybrook.edu/about-us/News/…

@jakejsullivan: That’s just one example of many where getting technologists and strategists together to really go through the underlying assumptions of where all this is going is important. Most of this conversation floats above at a plane of abstraction. As a result, we pursue a policy with a lot of hidden assumptions that we haven’t fully validated or unearthed. That’s my main concern and something I’m turning over in my head and thinking about how to articulate better than I just did.

@jakejsullivan: A world in which we are rapidly approaching AGI and ASI versus a world in which it’s a boundary and jagged pathway towards greater capability — what you would do across a range of different inputs is different in those two worlds. Yet I don’t think we sit and work out what our relative confidence is in one world as opposed to the other. What are the can’t-fail or the must-dos, the no-regrets moves? Then what are the things we have to be able to adjust as we gain more information?

My former boss @jakejsullivan was on with @jordanschneider at ChinaTalk a couple days ago, talking about AI policy, among other things. chinatalk.media/p/jake-sullivan Key insights on how the technical details matter: we need a better handle on the science of intelligence and theoretical foundations of AI.