Prof. Tanmoy Chakraborty, Professor at IIT Delhi gave a talk on Enough of Scaling Laws! Let’s Focus on Downscaling on 7 July.

The talk challenged the prevailing trend in Generative AI of building increasingly larger language models with ever-growing computational and data requirements. Prof. Chakraborty argued that scaling model size alone is neither sustainable nor sufficient for the next generation of AI systems. Instead, he proposed a future centered on smaller, specialized language models that excel at specific tasks and collaborate effectively to solve complex problems.

He presented recent advances in model compression, efficient adaptation, model orchestration, and collaborative reasoning, demonstrating how carefully designed small models can often match or even outperform much larger systems while being more efficient, interpretable, and easier to deploy. He envisioned an AI ecosystem where intelligence emerges from the coordinated interaction of multiple specialized models rather than a single monolithic system.

Prof. Chakraborty holds the Rajiv Khemani Young Faculty Chair in AI and is a Professor in the Department of Electrical Engineering and the Yardi School of Artificial Intelligence at IIT Delhi. He leads the LCS2 research group, where his work focuses on developing economical, interpretable, and adaptable language models. A recipient of several prestigious national and international honors, he serves as the Editor-in-Chief of ACL Rolling Review and has held key leadership roles in the NLP community, including Program Chair of EMNLP 2025 and Local Organising Chair of AACL 2025. He is also the author of the textbooks Social Network Analysis and Introduction to Large Language Models.

For more details: https://tanmoychak.com/

July 2026