We are used to compilers/linters/formatters of our beloved programming language. In the era of neural net (AI) we want to get the best of the world and make reviews more machinery and more "smart". How do we achieve that?
We are programmers, right? We all came from a deterministic, logical world. With LLMs we lose determinism. Decision models talk in typed structures instead.
Everyone argues about which neural net writes better code, benchmarks this, benchmarks that. Meanwhile the argument itself costs more than just picking one and using it.
Today is definitely not the best time for open-source contributions. There are multiple questions that can't be answered freely, but the main one is: "Why do this anymore?".
A payment handler resolves its method ids into class-level constants. One day charges start landing with a NULL method and the only fix is a pod restart. Why? This is a real production bug that makes an excellent interview question.
We expected AI to boost our productivity 1000x with new possibilities, but we know that it is just a fairytale. Now everyone from housekeepers to presidents has access to AI on the web like ChatGPT, and we have to deal with that.
I send out a monthly email summarizing new posts and insights on software engineering, developer life, and programmer philosophy. If you’d like to stay updated, drop your name and email below — I’d love to keep you in the loop.