• 0 Posts
  • 20 Comments
Joined 1 year ago
cake
Cake day: June 10th, 2023

help-circle












  • Which is doubly worse because those higher-level libraries are black boxes, and you can’t always fix things that arise inside of them easily.

    If by “higher level” you mean something like Java libraries, I’d say the opposite is true - at least if you don’t have the source for a Java class it is trivial to decompile and have something immediately readable. Can’t say the same for something like a dll originally written in C++.




    • old webpages (like from the 2000s) are fast and snappy
    • new webpages take much longer to load

    Modern webpages are less like a page and more like a full blown application. If you’re not careful you’ll get an unoptimized mess, which is exacerbated when you put a bunch of ads on top.

    That being said I don’t have memories of everything being snappy 20 years ago - there were messy scripts and gigantic images that made Geocities and Angelfire sites near unusable back then as well.


  • Getting them to do anything other than spit up probabilistically correct information and sentences would require them to have comprehension, not just associations between strings of words.

    I would argue that associations between pieces of information is the cornerstone of how you form comprehension. While I agree that the technology is somewhat overhyped, the fact that you can instruct models to perform arbitrary tasks that the model was not specifically trained to do is evidence of some level of “comprehension.”