In a Hindustan Times interview, Grandmaster Vidit Gujrathi discusses his dissatisfaction with his current chess standing, unveils his AI tool ‘Kibitz’, and outlines the broader implications for the sport and technology integration.
Based in Bengaluru, 31‑year‑old Grandmaster Vidit Gujrathi has recently launched an AI‑driven chess tool called Kibitz, promising to reshape commentary and analysis in professional chess. In a candid interview he revealed two earlier projects – a web‑app for blindfold chess and a puzzle‑solver built from real‑tournament positions – before diving into his latest venture.
Historical Context of AI in Chess
Chess players have been early adopters of artificial intelligence. The watershed moment arrived in 1997 when IBM’s Deep Blue defeated Garry Kasparov, marking the first time a machine out‑performed a world champion. Since then, successive generations of engines and, most recently, large language models (LLMs) such as GPT‑4 have pushed the boundaries of move prediction and strategic insight.
Kibitz: Modeling a Grandmaster’s Mind
Gujrathi explained that a call for entries on NVIDIA’s X platform sparked the idea for Kibitz. He wanted a system that went beyond raw engine evaluation and captured the “human side” of a grandmaster’s decision‑making. The prototype currently predicts the first move with 56 % accuracy, and the first three moves with 86 % accuracy – numbers that already outperform many conventional engines in broadcast settings.
Self‑Taught Technical Journey
His technical fluency is self‑directed: he consumes blogs, watches expert YouTube channels, and reads research papers. Influences include OpenAI co‑founder Andrej Karpathy’s tutorials and Y Combinator founder Paul Graham’s essays on startup thinking. By following top AI and ML thought leaders on X (formerly Twitter), Gujrathi continuously refines his toolkit.
Future Prospects and Impact
Testing Kibitz on recent Grand Chess Tour rapid and blitz events yielded a 60 % first‑move success rate. Gujrathi aims to scale the model with larger datasets to improve reliability and eventually embed AI commentary into live broadcasts. If successful, viewers could gain deeper strategic insight, while players might use the tool for preparation, potentially reshaping the competitive ecosystem.
