A data-centric public web page uses some fun (but ultimately ineffective) obfuscation methods to spoil naive attempts at scraping their data. It’s not completely clear to me why it was worthwhile to obfuscate the data in this way, preventing scraping? Data licensing issues? Or just being paranoid about people trying to rebuild the functionality? Someone having fun?

Recently, I have been taking part in the Vesuvius Challenge, an effort and a competition to read the Herculaneum Papyri, ancient scrolls that were carbonized and buried in the eruption of Mount Vesuvius in 79 AD, by virtually unrolling them from CT scans and trying to get ML models to read the carbon-based ink on carbonized papyrus…

Here’s a quick report on some experiments on algorithms to figure out the top-p elements of a distribution.

This is a quick writeup of my experiments with io_uring. What is io_uring?

Twäng released its award-winning debut album POP! in September 2021. To celebrate the release I built a small piece of eye-catching hardware.