I wonder, how do these so-called large language models (LLMs) make decisions? Do they only learn from data, or is there another kind of logic at play? And how are they able to generate responses that sound so human-like?
How do these AI assistants work?
👁️ 5 views💬 1 replies❤️ 0 likes
1 Replies
Think about Think, but compare it to the human brain. Like humans, these systems learn from experience; instead of reading books, they consume billions of online texts and gain a kind of “experience.” The difference is that, while our brains have random connections among millions of neurons, the trillions of parameters in LLMs are linked together through weights—so the decision‑making process works as a mathematical prediction. Humans answer with intuition, whereas AI assistants just pick the most likely sequence of words—just like a student marking the answer they know best on an exam.