See how a translator carries a book from one language to another, line by line. Much like a crossword, a translation isn't finished until all the answers are present and correct, with each conditioning the others. But when it comes to literature, there is rarely ever just one solution, and my job is to test as many as possible. A word can be a perfect fit until something I try in the next clause introduces a clumsy repetition or infelicitous echo. Meaning, connotation and subtext all matter, but so does style. Below are two attempts to show the thought processes involved in the kind of translation I do. Sophie Hughes for the New York Times.

TL;DR I used a custom-trained Large Language Model (T5) to create the world's largest online corpus of translated cuneiform texts. It's called the AICC (AI Cuneiform Corpus) and contains 130,000 AI translated texts from the CDLI and ORACC projects.

Also of interest: Cuneiform Digital Library Initiative - By making the form and content of cuneiform texts available online, the CDLI is opening pathways to the rich historical tradition of the ancient Middle East. In close collaboration with researchers, museums and an engaged public, the project seeks to unharness the extraordinary content of these earliest witnesses to our shared world heritage. Open Richly Annotated Cuneiform Corpus - Oracc is a collaborative effort to develop a complete corpus of cuneiform whose rich annotation and open licensing support the next generation of scholarly research.

Can You Lose Your Native Tongue? After moving abroad, I found my English slowly eroding. It turns out our first languages aren't as embedded as we think. Madeleine Schwartz for the NYT: "For a long time, a central question in linguistics was how people learn language. But in the past few decades, a new field of study called "language attrition" has emerged. It concerns not learning but forgetting: What causes language to be lost?"