> I read a vanishingly small amount of the actual code.
This sentiment kind of saddens me. I'm all for burning tokens to write throwaway code just to prototype a solution, but I don't get not reading (or at least familiarizing yourself with) the code that you will deploy to prod.
maybe they're betting on improvement in models to plateau, and that having a fairly stablized capable model that is orders of magnitude faster than running on GPU's can be valuable in the future?
Sure, but, in the real world, for the software to deliver a solution, it doesn't really matter if something is modelled in beautiful objects and concise packages, or if it's written in one big method. So for those that are more on the making /things/ side of the spectrum, I guess they wouldn't care if the LLM outputs code that has each iteration written separately.
It's just that if you really like to work on your craftsmanship, you spend most of the time rewriting/remodelling because that's where the fun is if you're more on the /making/ things side of the spectrum, and LLMs don't really assist in that part (yet?). Maybe LLMs could be used to discuss ways to model a problem space?
I think this is showing the difference between people who like to /make/ things and those that like to make /things/. People that write software because they see a solution for a problem that can be fixed with software seem to benefit the most of LLM technology. It's almost the inverse for the people that write software because they like the process of writing software.
If anyone is interested in the cryptography mistake that Sony made I recommend watching the Console Hacking talk at 27c3 by the fail0verflow team: https://youtu.be/DUGGJpn2_zY?t=2096
Interesting to read that the complexity in his music is praised and seen as speaking to the intellect, whereas that is not the case when it comes to complexity in software.
I'm working on a new programming language and every point resonates with me on a spiritual level. One thing that helped me a lot was writing more integration tests at the outer edges of the application, opposed to unit tests of individual modules. Usually the logic doesn't change between rewrites, so keeping a somewhat stable interface keeps me from breaking stuff in between grand rewrites.
It looks like they request these permissions: user-top-read, playlist-read-private, playlist-read-collaborative and they also gets your your name, username and profile picture.
This sentiment kind of saddens me. I'm all for burning tokens to write throwaway code just to prototype a solution, but I don't get not reading (or at least familiarizing yourself with) the code that you will deploy to prod.