• 0 Posts
  • 220 Comments
Joined 3 年前
cake
Cake day: 2023年6月17日

help-circle
  • The problem is, YouTube has no real competition. No one has the same thing they do. And the same applies to a lot of things today.

    Another issue is, subscription services can just raise the price and then start charging people more without those people doing anything, or possibly even noticing, which differs from individual purchases where you had to make a judgement about the price each time. Now, people have to make an actively cancel in order to not agree to new prices (or EULA changes for that matter). It should really be the other way around, so that if a service raises their price, people have to actively agree to the new price in order for the service to keep charging them.

    Can’t speak for Uber Eats or Starbucks though.










  • It’s not though. Programming languages, like assembly before them, are deterministic. If you run the same C code again the same environment, it will do the same thing, and altering the code will alter the behaviour correspondingly. It’s possible to reason about it. The same does not apply to LLMs. You can’t reason about their behavior, when means you can’t build anything non-trivial with them. All that is mentioned in the article.







  • I know this is humor, but for the record this wouldn’t work. Each simultaneous TCP connection needs a unique four-tuple (source address, source port, destination address, destination port). If a lot the people behind the NAT try to connect to the same place (destination address and port) at the same time (something popular like Google, YouTube or Netflix), and their source address is the same, the source port needs to be different for each connection. So after at most 65535 connections within a short time the NAT would run out of ports and no one behind the same NAT would be able to open new connections to the same place until the NAT mapping expiries.

    So you could have at most tens of thousands of people behind the same NAT, maybe even fewer to make it reliable.





  • If you’re running Wine on a case-sensitive file system, and you it tries to open a file, it would first try to open a file whose case matches exactly. But if it doesn’t find one, it would then need to list all the files in the directory, normalize their case, and go through them all to see if there is a file with the given name but in a different case. That can take some time if there is a lot of files in the directory.

    But if you’re on a case-insensitive filesystem, the FS can keep case-normalized names of all files on disk, so you can do a case-insensitive open just as fast as you can do a case-sensitive open.

    BTW, another application that can benefit from this is Samba, since SMB is case-insensitive.