Operation is the bottleneck

Operations are also assisted by AI, but I didn't make much progress in my inexperienced Azure operations experiment with maximum AI support for me.

It is difficult without experience to identify what is wrong when it does not work as expected because it only works when various factors are intertwined.

Either the container is not being built, or the build is not being pushed, or it is being pushed but the server continues to use the old image, or the log output is not supplemental.

You can't do it with no experience in this act itself, which you just enumerated, "These are the elements, right?"

I can do it because experience has increased my resolution.

I asked the AI and it didn't sound good.

There may have been a better way to ask, but since I haven't been able to do it, I think it requires a much more difficult ingenuity.

The cognitive resolution of the person observing the problem is low, so they ask questions with vague problem descriptions.

Often distorted cognition, overlooking facts that are important to solving the problem.

The AI responds by drawing from the vague description, but because the input is vague, the response is also vague, and even if you force it to explain specifically, it will not be the correct answer, but the most common one.

These operational problems and their resolution are not in the online text.

Costly to isolate and disclose because in most cases it is tightly coupled to the circumstances of a particular organization

There is not a high incentive to write stories about problems that only occur in limited circumstances.

You need to observe multiple cases to get a good abstraction, but most people only touch one environment.

---