Julia Turc

AI researcher and content creator specializing in efficient machine learning techniques, particularly around model compression, training optimization, and world models.

• Video: How does 4bit quantisation work (2026-04-14) - Discusses evolution of training large-language-models with reduced precision, focusing on 4-bit floating-point (FP4) training challenges. Highlights cost of training LLMs (Gemini Ultra: ~$191Min2023,[[entities/gpt-4|GPT-4]]:$78M in 2023) and shift toward Quantisation techniques to reduce computational demands.
• Key insight: Reducing precision from 16-bit to 4-bit significantly lowers training costs while maintaining model performance through advanced Quantisation methods.

2026 04 14 How does 4bit quantisation work

• 2026-04-10 [2026-04-10-TurboQuant-Reducing-LLM-Memory-Foo

• Video: World Models in AI: Concept, Implementations, and Applications (2026-05-24) - Comprehensive overview of world models in AI, defining the concept, exploring implementations, and highlighting applications. See World Models in AI: Concept, Implementations, and Applications for detailed notes.