Jordan Tigani, renowned “database person” and duck herder/CEO at MotherDuck, gives us the gift of “The Analytical Language of John Wilkins” by Jorge Luis Borges. We talk about about the potential of language, the limits of language, compression, sloppy ontologies, LLMs, what thing the universe is, simulated annealing, our vague comprehension of what embeddings are, and why it’s unfortunate that there’s no way to not sound pretentious when talking about Borges.

A few notes and links related to the conversation:

• Read “The Analytical Language of John Wilkins”. It’s quick!
• The Wikipedia article about the essay is pretty good.
• Jordan is also a very talented writer about data and technology, which is great because he’s better informed than almost anyone on the planet. Read Big Data is Dead from the MotherDuck blog.
• Simulated annealing on Wikipedia 🤯
• Trie
• Huffman coding
• Linnaean taxonomy
• Why Fish Don’t Exist by Lulu Miller
• The Importance of Common Data Schemas and Identifiers from the Cloud-Native Geospatial Forum blog (written by Chris Holmes and me)
• Brad DeLong finds that ChatGPT doesn’t know what ISBN numbers are
• Edgar Codd, proto database person
• WorldCat gets scraped (Hacker News discussion)

I asked ChatGPT to write a limerick about Borges and I wish I didn’t have to admit that it’s pretty good:

🍀🇦🇷🍀

There once was a poet named Borges,
Whose stories were full of deep forges,
In labyrinths vast,
Where time could not last,
And mirrors reflected strange Georges.

🍀🇦🇷🍀