• 101 Posts
  • 282 Comments
Joined 2 years ago
cake
Cake day: April 4th, 2025

help-circle









  • Completely honest opinion here as a software developer, if you don’t use AI in coding now, you are probably not working effectively. This is because really understanding and reviewing code not written by oneself requires a lot of time, and GenAI can (so far) not selve the very difficult problem of keeping a large codebase with a lot of changes and drift in requirements consistent.

    If you do use GenAI to generate code, it is extremely hard to both review the output thoroughly, and be faster than writing it by hand.

    For some cases, like code auditing, it might not matter. But for much of software engineering, a substantial part of it is a continuous learning process - learning requirements, constraints, specs, what the code should do, its structure, and so on.

    And GenAI short-circuits that - you invariable understand less of the code. This is not different from a student of physics or high-frequeny electrical engineering which does his or her homework on Maxwell’s equations with ChatGPT: They might get a result quickly, but at the cost of really understanding the matter.

    Again, there might be cases where this does not matter. For a throw-away web page or a script used once, it might be good enough. But in most cases, understanding what you do does matter.












  • Germany has 80 million people, China 1.400 millions. The per-capita consumption of Germany is still much higher. Chinas energy production is rapidly growing, and the share of renewable energy in it is growing as well.

    Moreover, Germany is the largest economy in the world’s largest democratic economic bloc, since the US objectively ceased to be democratic. So, Germany and the EU have the duty to show that they can do better. “We had democratic governments but they had to serve the short-term interest of capitalist megscorporations” will not sound like a good excuse to our grandchilds when they have to live on a devastated earth.










  • He is right in many aspects. However, for example Guile, Racket or SBCL Lisp are a lot faster than Python, and they are more powerful than Python while being much simpler, which makes them very interesting for prototyping algorithms.

    The ideal combo for such things would be a powerful interactive language with debugging capabilities combined with Rust, perhaps even a language similar to Rust. I do not know of any easy solution for that, but it could be made work with Rust’s macro system (since the above languages all provide C ABI FFI).

    Also,most of design happens when a program is written first. It is hard to re-design a larger program later, which might be one reason to re-write it. So maybe the main reason for rewriting is to get a better design, not the faster language - the faster language might be an additional reason.

    Also, LLMs can produce terribly bloated Rust code - not only by a factor ten larger, but also much slower.

    Also, CRUD apps are uninteresting to many programming enthusiasts, and for companies, optimizing them is only interesting if it saves them money.