2. Modern LLMs like GPT take incredibly powerful hardware to run. Even if offline options were available, useres would need powerful graphics cards to get quick and high-quality responses