Longreads + Open Thread

Cantor, Stripe, PE, AI and Software, English, Pair Trades, Crypto, Maps

Byrne Hobart Byrne Hobart

Longreads

You're on the free list for The Diff. This week, paying subscribers got a look at the case for letting the Fed control fiscal policy, too, and how that would work ($), why we'll be creating many more sources of clean, structured data even though LLMs can handle messy, unstructured stuff ($), and what the world will look like when everyone has more access to the average level of intelligence and social skills ($). Upgrade today for full access.

Upgrade Today

Books

This Earthly Globe: A Venetian Geographer and the Quest to Map the World: For all their odd little failure modes, it's sometimes striking to think that large language models have such a comprehensive model of the world, created entirely from identifying and generalizing statistical patterns in text. But, in a way, it's a return to a very old tradition. In 1550, a Venetian publisher and public servant produced the first volume of Navigationi et Viaggi ("Navigations and Voyages"), the first comprehensive book about what you could see and where you could go throughout the entire world—it had accounts of trading voyages to Africa, it revived a more accurate translation of Marco Polo's memoir, details about what was happening back in Muscovy, and descriptions of North and South America.

You can't coherently write a history of geography, because every inhabited place has been implicitly mapped by whoever lives there. What you can write is a history of geography as it's practiced from one particular location: what would a 12-century Venetian think you'd see if you started in Constantinople and traveled east, and how would a 16-century traveler's view differ?

From a European perspective, these maps were gradually clouded by religious conflict. The rise of Islam put hostile land and naval powers between Europe and Asia, which meant that people were using some very old sources (Ptolemy, Herodotus) and otherwise relying on rumors. It's not as if the Ottoman Empire had a strong interest in sharing detailed maps and travel advice with their biggest rival in the Mediterranean.

They also had to deal with what's best described as an early example of a political deepfake: in the 12th century, the Eastern Roman Empire received a letter purporting to be from Prester John, who claimed to rule a vast and wealthy Christian kingdom somewhere in Central Asia. This letter got mixed up with news of the actual battle of Qatwan, in which the Seljuk Turks were defeated by Qara Khitai. Turks being known to be Muslim, and Qara Khitai being unfamiliar to Europeans, it didn't take much motivated reasoning for them to conclude that Prester John was real, and that if they launched yet another crusade, he'd surely show up with reinforcements. Which didn't happen, though a few centuries later the narrative had shifted from Central Asia to Africa, and European explorers did stumble on Ethiopia, which had been a Christian country since the 4th century.

Navigationi et Viaggi collected narratives from people who had often led more interesting lives than they personally preferred. al-Ḥasan ibn Muḥammad ibn Aḥmad al-Wazzān al-Zayyātī al-Fasī, for example, grew up in Muslim Granada, grew up in Fez, was kidnapped by Spanish pirates, and kept under house arrest by the Pope. He wrote extensively about North Africa (as it turns out, 16th century Fez had a lively red-light district). He eventually converted to Christianity, at least nominally, and took the name Leo Africanus. Marco Polo had a similarly tricky experience, where he traveled throughout Asia taking notes and striking deals, and ended up at the court of Kublai Khan, where he was simply too interesting and helpful to be given permission to get home. He did eventually manage to find his way home, and then ended up in another prison, in Genoa, where he collaborated with a novelist to publish his memoirs.

So the task for Giovanni Battista Ramusio, the author of the book, was to take an enormous store of information—memoirs, journals, private letters secretly leaked or stolen, nth-hand rumors, works that had gone through multiple rounds of translation, political screeds, stories written under varying kinds of duress, etc.—and weld them into a coherent narrative. Which is, interestingly enough, a task that starts out impossible but actually gets easier as the story gets more comprehensive. If Marco Polo's stories line up with other narratives about Central Asia, and if the hypothetical kingdom of Prester John was simply too big to miss, it had to be cut from the narrative. If the rhubarb described in a story about its point of origin sounds like the rhubarb that shows up in Venice, it's probably real. As you collect more comprehensive information about the world, even if it's mostly text, you get to a point where an internally-consistent story takes shape. And this book is a nice meta-book about how that process works. It’s a portrait of a very weird world, but one that’s in the process of becoming recognizable to us.

Open Thread

Leave a Comment

Diff Jobs

Companies in the Diff network are actively looking for talent. See a sampling of current open roles below:

Even if you don't see an exact match for your skills and interests right now, we're happy to talk early so we can let you know if a good opportunity comes up.

Find a Role

If you’re at a company that's looking for talent, we should talk! Diff Jobs works with companies across fintech, hard tech, consumer software, enterprise software, and other areas—any company where finding unusually effective people is a top priority.

Find Talent with Diff Jobs