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

Researchers propose MIPO, a self-supervised framework that improves LLM personalization by 3-40% without additional labeled data

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

Research reveals LLMs can 'self-purify' against poisoned data in RAG systems, identifying and down-ranking falsehoods