Unless a Google Youtube engineer with knowledge of the id creation mechanism chimes in regarding Youtube ids, I still think that there is some information embedded in the 64 bits. They could easily scramble the bits around to encode time and other data -- sort of like Twitter's
Conversation
Replying to
Snowflake algorithm but a more obfuscated version of it. Twitter's snowflake algo simply embeds the timestamp in the higher order bits and the server and datacenter ids in the lower order bits. What I'm wondering is what it would look like if one were to embed a timestamp but
1
scramble the bits around and mix in other metadata in other bits. After base64 encoding (which Youtube essentially does to their 64 bit ids), it would appear that each character has even distribution if you examined a large sample of ids. If they're scrambling bits, the first