A Parrot That Reads Nietzsche
Don't get me wrong, I understand why people turn to AI for advice - it feels like we’re talking to something smart. The branding even leans into it - “Artificial Intelligence.” These systems are smooth. They can draft poems, summarize legal cases, even mimic banter. They feel like conversation.
But feeling smart isn’t the same as being smart. Parrots can talk too. They can say “hello,” answer questions they’ve been taught, even surprise us with how human they sound. But we don’t ask parrots how to handle our taxes or whether we should break up with someone. We know they’re mimicking, not thinking.
Large language models are the same thing, just cranked up to a galactic scale. They’re parrots trained on the entire internet. Ask them a question and they’ll give you something plausible, polished, even convincing. But there’s no mind behind the curtain. Just probability math spitting out the most likely next word.
We Didn’t Make Computers Smarter, We Just Made a Language to Math Converter
This is the key point: the breakthroughs of the last few years weren’t about making computers “think.” They were about teaching computers to handle human language.
When you type a question into ChatGPT, you’re not having a chat with a computer like you would with a friend. You’re feeding it a string of words, Your words then get turned into numbers, those numbers get crunched through billions of weighted connections, and the output is another set of numbers that gets converted back into words.
It’s not thought. It’s pure math. You are communicating with a giant, word-shaped calculator that happens to be really good at sounding friendly.
And that’s useful! Calculators are great. Predictive text is great. But let’s not pretend it’s more than it is. This isn’t your mentor, your therapist, or your life coach. It’s autocomplete on steroids.
Why It Matters
If you understand this, you can use AI responsibly - as a tool, not a guru. Let it help you draft an email, brainstorm a campaign, or summarize a dense report. But don’t let it hand you life advice or moral guidance. Because at the end of the day, you’re not talking to an intelligence. You’re talking to a parrot with access to a trillion words.
And here’s the part OpenAI and Google don't like to admit: no matter how much better we make these current versions of AI, they’ll never evolve into true Artificial General Intelligence because they are built on a shortcut. AGI isn’t hiding in the next model release, waiting to pop out once we add enough training data. This tech won’t suddenly “wake up.” If AGI ever arrives, it will come from a completely different approach - from someone building a different kind of machine, not from squeezing more juice out of the math trick we’re using now.
So go ahead and marvel at how good the parrot sounds. Just don’t mistake mimicry for mind. ok, Jar-Jar?
