Here I was hoping that if you took the UTF-8 representation in bytes and decoded it as ASCII, you would get something interesting. But no, just Unicode characters. Almost interesting is that none of the bytes are valid ASCII characters (< 128), which you might expect for the first byte of every UTF-8 codepoint due to backwards compatibility for ASCII encoding, but perhaps not for the subsequent bytes that comprise the rest of the grapheme.

I’m finally starting to understand the appeal of numerology.

These formats are generally not very compressible by general purpose compression algorithms, as they are already compressed formats themselves, each with a compression algorithm specially tailored to their content type.

Been awhile since I’ve looked, but you might consider pixz:

https://github.com/vasi/pixz

Self correction, the author is a different Elizabeth Anderson, also a philosopher.

Fair enough. Doesn’t bode well for DoH in authoritarian regimes.

If you like this, the author (Ellie Anderson) also has a podcast with her co-host David Peña-Guzmán called Overthink.

https://www.overthinkpodcast.com/aboutus

Even Palo Alto notes that they can only effectively block DoH if you’re MITMing all https traffic already (e.g. using a root certificate on corporate-managed devices). If not able to MITM the connection, it will still try to block popular DoH providers, though.

https://live.paloaltonetworks.com/t5/blogs/protecting-organizations-in-a-world-of-doh-and-dot/ba-p/313171