The problem is the way this is written and formulated, these red flags "this thing has been thought and written extensively with an LLM" make the whole (rather compelling at first consideration but hollow-feeling when thinking more about it) thesis not very enticing.

I always find it crazy how much bullshitting can go around processes. Whether it's bullshitting about doing agile "right", or adapting to AIgile and do it "right" again.

The truth is that when you have a group of people building something together, they need to communicate. The larger the group, the harder it becomes. Processes are here to formalise that communication. "You have to declare X" is a result of "in the past, some of the people involved were not informed of X and it was causing issues big enough that we added a rule".

The problem with methodologies is that those who need them are those who lack experience (because experienced people will know how to build their own processes), and by definition those who lack experience will tend to "cargo cult" (they cannot know, they haven't experienced it yet).

And there is apparently a big market in selling bullshit consulting to inexperienced people.

AI is a tool that may make developers more productives, but it won't remain the fact that software is built by a group of people (from software devs to sales through product and testing) who need to communicate, period.