A Java Geek weekly 129

I wonder how long it will take to change organizations to actually trust the AI to review the code. The ones who did already had some nasty surprises. And yet, is there any other alternative.

I am of two minds. It’s a cool engineering feat, but does it rewards the efforts the vendor put into creating the product in the first place?

I was so ready to start learning Gleam, the language that would introduce me to the BEAM. Then, coding assistants started to become useful.

Mathematically proving the code is a great way to finally trust the coding assistants! However, the token consumption seems to bit on the high-end. I’m also wondering whether you need to use a formally-provable language, such as Coq or Idriss.

Most of us, including me, forgot that we are only allowed to work in abstractions because the world was more or less stable. I was first introduced to the "real" world when working on a worldwide project. In July 2014, Russia introduced a law that mandated that any company collecting or processing personal data of Russian citizens must store that data on servers physically located in Russia. My colleagues rushed to split the relevant data from the database and to develop the cheapest software possible to acces it. The business already had decided the country was lost. Unfortunately, they were right in so many ways.

I actually am super curious about this.

In the API realm, the decoupling layer between client and server known as API Gateway unlocks a lot of options. AI Gateways do the same in the realm of AI.

Philosophical. Only future can tell.