Oh No, It's Not a calculator, But it's still not AI

Published on: November 29, 2024

The way LLMs work is amazing, it's not a calculator at all, but it's still not intelligence, or is it really. We are still trying to find what it is. Currently, it's a blackbox, just like our brains. We know our brain has signals passing, some material or electrical, but how does consciousness arrive is something we still don't know.

No one knows exactly what intelligence is yet, and this is something scientists and researchers agree on. Intelligence can mean solving problems, learning, being creative, or understanding emotions, but we still can’t figure out what connects all these abilities. Even in our own brains, we don’t fully understand how signals and neurons create the experience of thinking or knowing.

For now, what we call AI is more like a skilled predictor. It spots patterns and suggests what might come next, but it doesn’t truly understand anything like we do.

It’s kind of like Nostradamus, whose predictions seem amazing only when we match them to something that has already happened, like calling a fire in the sky a plane crash or a meteor. LLMs do the same, they say something that sounds smart, and we decide how it fits.

You can also think of LLMs like pareidolia, the way we see faces in clouds or hear words in random noise. They generate outputs based on patterns, and we find meaning in them, even though the meaning isn’t really there.

If true intelligence means understanding and being aware, we’re not there yet. These machines don’t think or feel, they just process data in ways we’ve taught them to. Without knowing how our own minds turn information into thoughts and ideas, we can’t create something that truly thinks like us.

For now, LLMs are tools that mirror the knowledge we’ve fed them. They seem smart because we see meaning in their responses, but that’s a reflection of our own intelligence, not theirs. What makes them fascinating is how they remind us of the questions we still haven’t answered, about intelligence, consciousness, and what it really means to think.

We teach the machine,
It repeats what it has learned.
Thinking is still ours.
Made with Mevin.ai