"I'm stressed about my job."
"What's making you feel stressed about your job?"
"Well, my boss doesn't seem to value my work."
"Tell me more about your boss."
If you were messing around with computers in the late '70s or early '80s, that dialogue probably just gave you a hit of nostalgia.
Way before ChatGPT was drafting your emails or Claude was summarizing huge PDFs, we had ELIZA. It wasn't flashy — just green text on a dark screen — but for a lot of us, it was the first time a machine actually felt like it was talking back.
ELIZA was built at MIT in the '60s by Joseph Weizenbaum and was one of the world's first chatbots. The version most people remember, DOCTOR, acted like a therapist by basically mirroring whatever you said back as a question.
When
I was 12, I spent a whole week painstakingly typing ELIZA code from a magazine into my Commodore VIC-20. I was so pumped when it finally worked that I spent hours chatting with it, half-hoping it could help me navigate the drama of middle school in the '80s. I clearly remember asking ELIZA if I should drop my savings on a new Atari game or a Rubik's Cube.
I knew deep down it wasn't actually "listening," it was just flipping my sentences into follow-up questions. But after a few minutes, I'd totally forget there wasn't a real person on the other side.
Compared to the powerhouse models we have today, ELIZA had less brainpower than your microwave. But for a kid discovering tech, it felt like a sneak peek at the future.
I'd sit there at that monochrome monitor, testing its limits. Could I trip it up? Did it have a sense of humor? Would it remember what we talked about five minutes ago? Every reply made me want to experiment more.
Of course, ELIZA wasn't exactly a genius. But at the time, it felt like staring through a crack in the door — a real, tangible hint that artificial intelligence was going to be much more than just science fiction.
Fast forward to today, and we've gone from a simple bot that just repeats your words to large language models that can audit contracts, code software and help kids with calculus. It's an incredible jump.
Yet, despite that massive leap in technical complexity, the core lesson of ELIZA hasn’t changed a bit.
It turns out that even as the code gets smarter, our human reactions stay remarkably the same.
Even back then, people treated ELIZA like it actually understood them. Weizenbaum was famously unsettled by how quickly users bonded with what was really just a few clever coding tricks.
Our current AI tools are obviously on another level; they can actually do the heavy lifting. But we still have that same human habit of assuming a smooth conversation implies a human-like mind, even when we know it's just sophisticated math.
So, how should we really be using this stuff? I don't think it's about handing over the keys to the castle. Think of these tools more like a super-speedy assistant that clears out the clutter. When they handle the repetitive tasks, you’re free to actually focus on the creative decisions, the stuff that really matters.
As wild as today's AI feels, I bet in twenty years we'll look back at ChatGPT the same way we look at ELIZA now. We'll laugh at the things it couldn't do and tell the next generation, "Man, you should've been there."
And they'll probably roll their eyes, just like we did.
But for those of us who remember that blinking cursor asking, "Why do you say that?", there will always be a soft spot for it.
In the end, perhaps the most "intelligent" thing about AI isn't the code itself, but its uncanny ability to act as a mirror, reflecting our own deep-seated desire to be heard.
JJ Rosen is the founder of Atiba, a custom software development firm, and Nashville IT support company. Visit Atiba.com for more info.
This article originally appeared on Nashville Tennessean: ELIZA felt like magic in the 1980s. It still explains AI's impact today
