3 ms·
The article is over sensationalizing something that really isn't that shocking. I thought they were going to index unlisted videos using brute force lookups of
by unreal37 8y ago
The article is over sensationalizing something that really isn't that shocking.
I thought they were going to index unlisted videos using brute force lookups of the URL, or that an API was leaking. Now that would be Hacker News worthy.
- userbinator 8y agoI thought they were going to index unlisted videos using brute force lookups of the URL If you're more interested in the format of YouTube video IDs, and whether it's possible to bruteforce them, see the last post at https://webapps.stackexchange.com/questions/54443/format-for-id-of-youtube-video https://webapps.stackexchange.com/questions/54443/format-for... tl;dr: they're 64-bit integers, base64 encoded, and randomly generated. I haven't been able to find any clear statistics on how many videos it contains, but it was slightly below 2^27 in 2008 (<80M) and based on the other statistic that between 2^18 and 2^19 videos are uploaded every day; in 10 years that's <2^12 days, so there's probably <2^31 videos on YouTube now --- roughly 2 billion. That means your chance of finding a valid video ID by bruteforce random guessing is less than 1 in 2^33. You'd probably be throttled/IP banned long before you got even a single valid ID.
- londons_explore 8y agoThey are not randomly generated. How could anyone ever prove they were random without reviewing the source code? Eg. It could be an auto-incrementing number encrypted with a reversible cypher.