In the labyrinth of Jorge Luis Borges’ work, themes of infinity, paradox, and the complexity of human thought are ever-present. Stories like “The Library of Babel” and “Tlön, Uqbar, Orbis Tertius” transport readers to vast, imagined realities and alternate worlds, subtly blending the real with the fantastical. The library represents the universe, and the books, human knowledge.

These themes resonate with the immense potential and complexity of large language models (LLMs), capable of generating a multitude of diverse and intricate responses. An LLM, in essence, can be seen as a modern reflection of Borges’ infinite library: a machine that attempts to make sense of the nonsensical, seeking coherence within a sea of endless possibilities.

If Borges were alive today, one can only imagine his thoughts on AI. Yet, invoking his name in this context feels somewhat misplaced. Borges had a unique ability to lead readers into fantastical realms without them even noticing, a subtlety that contrasts sharply with the overtly mechanical nature of LLMs.

An LLM resembles a search engine for Borges’ library, but instead of indexing topics, it finds comprehensible combinations within the vast expanse of human-generated text. Imagine a collection that not only contains all texts produced by humans but also encompasses everything that a human could potentially read and comprehend, even superficially. This aligns with Borges’ idea of expressing concepts that can’t be formed from existing notions.

Borges’ exploration of infinite possibilities continues to inspire and challenge our understanding, much like the capabilities of modern AI. In his work, we find a mirror reflecting the boundless horizons of AI, a testament to the enduring power of imagination.

Autobiography from The New Yorker.