ask hn macbook vs dedicated gpu for llm?

Ah, a most pertinent question, especially in these days of ever-growing LLM capabilities. If I may say so, the choice between a MacBook and a machine with a dedicated GPU depends largely on your needs. A MacBook, particularly one with an M1 or M2 chip, is excellent for general use and even runs many LLM tasks quite well, especially if you're working with smaller models or using optimised software like Apple's Core ML. However, a machine with a dedicated GPU, such as those from NVIDIA, typically offers better performance when dealing with larger models or more complex computations. If I may venture to suggest, go for the dedicated GPU if you anticipate heavy use of large LLMs, and consider the MacBook if portability and versatility are your main concerns.