The main solution to all the world’s problems is to make sure that each and every individual on the planet is smart, honest, empathetic, spiritual and wonderful in general. Yeah, it’s extremely hard to argue with this.
Someone Samuel Langhorne Clemens once said, “Truth is mighty and will prevail. There is nothing wrong with this, except that it ain’t so.”
In the meanwhile, we need to tackle issues with sloppy slope (or slippery slop, both sounds equally funny for a parody of an idiom) right away. But hey, were not developers mindlessly copying crap from StackOverflow, forums, even books a decade ago? Were not they (us) making mistakes leaking to merge requests simply because of a lack of competense, flying by deadlines or whatnot? How is a bad code produced by LLM differs from a bad code copied as is from StackOverlow?
a regular +500/-300 commit
Excuse me? I ruled out +200/-100 commits yesterday, and I am going to rule them out tomorrow. No matter who produced the code, the tasks/problems themselves should not have changed, should they? “Fix everything in this project” task is not something we shoud blame LLM for.
Of course. But who actually cares? Those are still pet projects. Let me quote a classic:
In the hands of people far removed from software development, large language models are doomed to two low-yield applications: you can amuse yourself generating memes of piglets drinking mojitos in the Kremlin, or you can reproduce a product that already exists on the market and that in its niche you’ll never catch up to. — Software Development in 2026
I see many people around alarmed by I don’t actually know what. They shift the responsibility to AI assistants and then they say, oh, the process has changed. The universal answer is the process mustn’t have been changed.
Reasonable tasks instead of “With the help of AI, make this leftpadding library to serve coffee and fly to the moon” should stay. Code review should stay. Rules on PR size and everything else should stay. Etc.
