How Much Better Are LLMs Than a Human?

“Humans are just models too.”

That’s what James Cameron recently said when talking about AI and copyright — and he’s kind of right.

We, as humans, consume vast amounts of information throughout our lives and remix it to create ideas, stories, and solutions. That sounds a lot like what large language models (LLMs) do too.

So how do we actually compare?

How Much Data Do Humans Really Use?

Take Shakespeare, for example. His entire vocabulary was around 20,000 words. A well-educated English speaker today might know 30,000–40,000 words. But more importantly, the average human brain can only actively hold a few chunks of information at once — and we forget things constantly.

We read books, watch videos, talk to people — and over time, our brains shape a sort of “model” based on repeated exposure, emotional context, and usefulness. But it’s slow, and extremely lossy. We don’t have perfect recall. We don’t have autocomplete. And we definitely don’t have 2025 Wikipedia in our heads.

What About ChatGPT (or Any LLM)?

ChatGPT was trained on trillions of words. Literally. From books, websites, academic journals, forums, social media, and more. It has a statistical understanding of how language works, how ideas connect, and it can draw from a staggering range of examples across domains.

While a human might remember the gist of a conversation from last week, ChatGPT can simulate the style of a 19th-century philosopher or generate Python code from scratch in seconds — all based on what it’s seen before.

It’s like having a virtual brain that’s read more than any human ever could, and never gets tired or bored.

So Where Does That Leave Us?

I use ChatGPT as a scaffold — a kind of brainstorming partner that helps me research, structure, and iterate on ideas faster.

Need a list of angles for a new article? I’ll ask it to suggest 10.

Trying to explain a tricky concept? I’ll use it to simplify and reword things until it clicks.

Want inspiration for a blog post? It helps me explore directions I hadn’t thought of.

I’m still the editor. I still inject the human vibe. But the speed and breadth it brings? That’s superhuman.

Final Thought: Models All the Way Down

If humans are models — shaped by the books we’ve read, the conversations we’ve had, the experiences we’ve lived — then LLMs aren’t so alien. They’re just faster, wider, and less forgetful versions of what we already do.

The big difference? We still bring emotion, values, and lived experience to the table. That’s our edge — and our responsibility.