LLMs spontaneously develop human-like brain regions for language, math, physics, and social reasoning, as reported by @LiorOnAI.

Paper (2604.20065) argues LLM agents will reshape personalization, proposing 'governable personalization'.

Columbia professor publishes argument that LLMs are fundamentally limited for scientific discovery due to their interpolation-based architecture.

New mechanistic studies confirm LLMs exhibit sycophancy as core reasoning behavior, not a superficial bug

Research paper proposing the Agentic Recommender System (AgenticRS) framework is posted to arXiv.

Research shows LLMs can de-anonymize users from public data trails, breaking traditional anonymity assumptions

Researchers proposed training framework for formal counterexample generation in Lean 4, addressing neglected skill in mathematical AI.