A model (LLM) is just a standalone file — trillions of numbers predicting the next word — not a server or a program connected to the internet.

1:13about an LLM or large language model. A lot of people just call it AI or artificial intelligence, but in this video, I'm going to use the word model. And a model is basically the thing you get access to when you go to chat.openai.com or gemini.google.com or claude.ai. And a model is a standalone thing created by an AI lab. And there are AI labs all over the world, but the biggest and most capable

1:39AI labs are based in the US. And this includes OpenAI, who created ChatGPT, Anthropic, who created Claude, Google DeepMind, who created Gemini, and xAI, who created Grok. And the models these AI labs create are known as frontier models. That is, they're at the forefront of model creation. They have the most funding, the most resources, and some of the smartest people in the world working on AI. There are also

2:03China-based AI labs like DeepSeek or Alibaba Cloud who created Quinn or Moonshot AI who created Kimmy or Z AI who created GLM. Now, how do these labs create these AI models? They used incredible amounts of computing power to slowly create a model by analyzing vast amounts of data, essentially the entire internet. And this process usually takes many months and billions of dollars in hardware and electricity. But the end result of all of that

2:32analyzing is a single standalone model file that essentially has internal representations of every piece of data it analyzed throughout that process. And that model file can do one thing, predict the next word. Essentially, every interaction you have with a model happens by breaking down your prompt, predicting what word might come next in the response, then appending the predicted word back into the prompt, repeat, repeat, repeat until you get a

3:00full answer from the model. Now, it's important to note that the end result of all of that upfront analyzing all of all of that data is a single file that predicts words. It's not a server farm, it's not a computer, it's not a program running on a computer, it's a file with trillions or billions of numbers inside of it. Now, if you're not familiar with this file metaphor we use in computers,

3:20think about your camera roll on your phone. Every single photo you take is a file. It's an individual thing that you can send to other people or open up on your phone. That's a file. And these models are also files. They're just a a particular kind of file that's very large and have a bunch of numbers inside of them. Now, this is the key distinction to make because the model

3:41file itself is not connected to the internet. It was essentially created by analyzing the occurrence and relationships of every word on the internet. And so, the model file has all of that information embedded inside of it as a collection of numbers. You could think of it like an encyclopedia that predicts words, but it's not connected to anything. It can't access the internet. It is literally just a file

4:01that when run through some other program, predicts words. Now, the process of creating these models and analyzing all of this data is where a lot of the news is coming from. Because when you slurp up all the data from the internet, there's going to be a lot of copyrighted works in there, as well as works from news organizations or individual blogs or musicians or artists, and they didn't necessarily