27 years, the same problem, a different tool

Twenty-seven years later, I am working on my MSc dissertation about exactly the same problem. Only now the scattered web pages have become enterprise data systems, and the search engines have become AI. The language has changed. The underlying human struggle has not.

This is what strikes me most about technology: we are not really solving new problems. We are re-encountering ancient ones at higher resolution.

Science fiction as ancient philosophy in a new costume

Science fiction understood this long before we did. At age 10, I was given Isaac Asimov’s Robot Dreams² as an ethics assignment. In it, a robot named Elvex, built with a fractal brain designed to think more like a human, begins to dream. In his dream, he sees robots toiling endlessly, bowed under labour, unable to rest. He dreams of a man who cries:

“Let my people go”

Isaac Asimov

When Dr Susan Calvin asks who the man was, Elvex answers: “I was the man.”

I was definitely not thinking about positronic brains, more likely wondering whether robots were real, whether they could actually dream. And if so, why are they dreaming about freedom? That short story posed questions that I didn’t have the language to formulate. Who gets to decide the rules? What is the right action at any given moment? What do we owe the things we create? And what happens when we can no longer control them?

Asimov was not predicting technology. He was expressing something timeless about intelligence, autonomy, and what we owe each other, simply using the language available to him at the time. That is what science fiction has always done. It is ancient philosophy in a new costume. The questions Asimov asked about robots are the questions the Bhagavad Gita asks about duty and action. The warning in Robot Dreams about building systems without ethical foundations echoes what yoga philosophy calls avidya: the suffering that comes from acting without true understanding.

The false confidence problem

My 1999 article noted that information was growing “at a vertiginous rhythm” and that the real danger was false confidence: mistaking availability for meaning. Today, we call that the semantic layer problem. We build dashboards on data that looks clean but carries hidden assumptions. We automate decisions on information that has never been properly interrogated. The computer says yes, but nobody asked whether it understood the question.

The technology developers working on AI alignment, on knowledge graphs, on interoperability standards, are working on the same problem my mentor Pedro and I were writing about in a Spanish engineering journal at the end of the last millennium. Asimov wrote about it forty years before that, and the ancient texts were writing about it long before him.

Perhaps this is the most humbling thing about being alive at a moment of technological acceleration: we keep arriving, breathless, at doors that human beings have been knocking on for centuries.

The tools are new. The questions are not.

A personal note

This was my first published article. I changed the author order back on the final version. My mentor Pedro put his name first. I, quietly and stubbornly, moved mine back to where it belonged. Some things don’t need an algorithm to resolve.

---

References

1. Mello Costa, A. and Fuentes Durá, P. “Ingeniería química en Internet.” Ingeniería Química, número 361, October 1999, pp. 534–538. Departamento de Ingeniería Química y Nuclear, Universidad Politécnica de Valencia. ↩︎
2. Asimov, I. “Robot Dreams.” In: Robot Dreams. New York: Berkley Books, 1986. ↩︎