AI and ChatGPT6 Sep 2025 15:55
Not what you are led to believe, as in we are told it is intelligent and will become super intelligent. This is drivel.
What is an LLM?
• A Large Language Model (LLM) is a type of AI trained on enormous amounts of text (books, articles, websites, code, etc.).
• It doesn’t “know” things in the way humans do, but it has learned the patterns of language: which words tend to follow others, how sentences are structured, and how ideas connect.
Think of it as a giant probability engine for words.
⸻
What is ChatGPT?
• ChatGPT is an application built on top of an LLM (like GPT-4 or GPT-5).
• It takes your input (the “prompt”), predicts the most likely useful next words, and keeps going until it’s finished an answer.
• On top of raw prediction, OpenAI adds instruction tuning and safety alignment so it doesn’t just spit random probabilities, but instead tries to be helpful, relevant, and safe.
⸻
Example: “The cat sat on the ___”
Here’s how an LLM approaches it:
1. It sees the input “The cat sat on the”.
2. From training, it knows the most common completions:
• “mat” (extremely common in children’s books)
• “sofa”
• “floor”
• “couch”
• “windowsill”
• even unusual ones like “spaceship” (from sci-fi stories).
3. Each option has a probability weight. “Mat” might be 70%, “sofa” 10%, “floor” 8%, “spaceship” 0.01%.
4. The model picks one, depending on how it’s tuned (more predictable vs. more creative).
That’s why sometimes ChatGPT sounds very “safe and normal” (choosing mat), and sometimes more creative (choosing spaceship).
⸻
Key takeaway
• LLM = word probability engine (predicts the most likely next text).
• ChatGPT = that engine + extra tuning so it’s conversational, helpful, and context-aware.
Sign In
Follow the stocks
