Learn how AI chatbots like ChatGPT and Claude really write, in about five minutes. Four short rides, one idea each. No account, no jargon.
A large language model, or LLM, is the engine inside AI chatbots like ChatGPT and Claude. It learned by reading a huge amount of text and practicing one job over and over: guess the next word. When you ask it something, it chops your words into small pieces called tokens, then writes an answer one token at a time, each time picking from a list of likely next pieces. That is why it can sound smart and smooth, and also why it can sound sure while being wrong. It is matching patterns, not looking facts up, unless it is given real documents to work from.
A large language model, or LLM, is a computer program trained on a huge amount of text so it can read and write like a person. Its core skill is predicting the next small piece of text, over and over. AI chatbots like ChatGPT and Claude are built on large language models.
First, your message is split into small pieces called tokens. Next, the model uses patterns it learned from its training text to work out the odds of every possible next token. It picks one, adds it to the answer, and repeats until the answer is done. It is predicting likely text, not looking facts up in a database, unless it is connected to search or given documents.
A token is a small chunk of text that an AI model reads and writes, often a whole word or part of a word. Common short words are usually one token, while long or rare words get split into several. As a rough guide, 100 tokens is about 75 English words, and AI services often set prices and limits by the token.
Temperature is a setting that controls how much randomness an AI uses when it picks the next word. A low temperature makes it stick to the most likely words, so answers are steady and repetitive. A high temperature lets it pick less likely words, so answers are more creative but also more likely to go off track.
AI models are built to produce likely-sounding text, not to check whether it is true. When the answer is not in what the model learned or in the text it was given, it may fill the gap with a confident guess. That made-up answer is called a hallucination, which is why you should check important facts and give the AI good source material.
A context window is the amount of text an AI model can see and use at one time, measured in tokens. It includes your question, any documents or notes you provide, and the conversation so far. Anything outside the context window is invisible to the model for that answer.
No. They are different technologies that solve different problems. AI language models predict and generate text based on patterns, while a blockchain is a shared record book that keeps entries from being secretly changed. In short, AI predicts and blockchain records.
Yes, it is completely free, with no account, sign-up or download. It runs in your web browser on a phone or computer and takes about five minutes. Everything is simulated right in the page with small word tables, so no real AI is called.