Transcribing Borges Lectures at Harvard

Diego Peralta

Diego Peralta / 2 months ago

4 min read––– views

Embracing the Art of Language and Technology in a Quest for Knowledge

In the realm of literature and philosophy, there are few names as revered as Jorge Luis Borges. Known for his mesmerizing prose and intricate storytelling, Borges captivated audiences with his profound lectures, leaving an indelible mark on the minds of those fortunate enough to attend. However, what happens when the echoes of his words begin to fade? How can we preserve and share his wisdom with future generations?

In this blog post, I embark on a remarkable journey, where language, technology, and passion intertwine. We'll explore the power of the OpenAI Whisper API, and how it became an invaluable tool in transcribing Borges' lectures at Harvard. So, fasten your seatbelts as we dive into this exciting adventure!

The Enchanting World of Jorge Luis Borges

Before we unravel the mysteries of technology and transcription, let's delve into the captivating realm of Jorge Luis Borges. Born in Argentina, Borges possessed a unique ability to weave intricate narratives that blurred the boundaries between reality and fiction. His lectures, often held at prestigious institutions like Harvard, offered profound insights into literature, metaphysics, and the human condition.

The Quest for Preservation: Transcribing with OpenAI Whisper API

Transcribing a lecture as profound and profound as Borges' requires a delicate balance of precision and artistry. Here, the OpenAI Whisper API emerges as an invaluable ally, empowering us to transcribe audio into text with remarkable accuracy. Whisper, built upon the powerful GPT-3.5 model, harnesses the potential of deep learning to understand spoken language and convert it into written form.

Let's take a closer look at the commands that demonstrates how to utilize Whisperr.cpp for transcription:

 git clone
 cd whisper.cpp && make
 ./models/ small
 ./main -m models/ggml-small.bin -f samples/jfk.wav

With just steps, we can unleash the power of the OpenAI Whisper and obtain a textual representation of Borges' profound words. Here is a sample of result:

00:00:00,000 --> 00:00:06,080
This Craft of Verse by Jorge Luis Borges. Following are the six lectures that

00:00:06,080 --> 00:00:11,640
Jorge Luis Borges delivered at Harvard University in the fall of 1967 and the

00:00:11,640 --> 00:00:17,880
spring of 1968. Born in 1899, he was by this time almost completely blind and

00:00:17,880 --> 00:00:22,440
thus addressed his audience without the aid of written notes. He learned his

00:00:22,440 --> 00:00:26,160
fluent English from his paternal grandmother, who had come to Buenos Aires

00:00:26,160 --> 00:00:31,680
from Staffordshire. His mother, a translator, and his father, a professor of

00:00:31,680 --> 00:00:37,120
psychology and modern languages, also knew English. These recordings, only

00:00:37,120 --> 00:00:41,400
lately rediscovered in the Harvard University archives, testify to his love

00:00:41,400 --> 00:00:45,720
of the language as well as his wit, erudition, and remarkable literary

00:00:45,720 --> 00:00:51,280
sensibility. Borges died in 1986.

And here is the link to the full transcript: This Craft of Verse by Jorge Luis .


With just the OpenAI Whisper API we have transcribed Jorge Luis Borges' lectures at Harvard, breathing new life into his profound words.

This adventure reminds us of the limitless possibilities that arise when we blend the artistry of language with the power of technology. As we continue to explore new frontiers, let us never forget the importance of preserving knowledge and sharing it with the world.

So, whether you're an avid literature enthusiast, a language lover, or a technology enthusiast, take inspiration from this tale and embark on your own quest to bridge the gaps between worlds. The power to transcribe and translate is at your fingertips, waiting to be unleashed.

Embrace the fusion of art and technology, and let your voice be heard across boundaries, cultures, and generations.

Subscribe to the newsletter

You can join here and I'll notify you whenever I publish a new post.

last issue