Context is eating prompt engineering, yes.
The mental model shift is: prompts tell the model what to do, context tells it what world it's operating in. You can have a beautifully crafted prompt and still get garbage output if the model doesn't understand the system it's working within.
What I've seen in practice — not just in code but in any AI-assisted workflow — is that the teams getting 10x results aren't the ones with the best prompts. They're the ones who've invested in structured context: clear SOPs, defined outputs, historical examples, constraint docs. The model just slots in.
The irony is that "context engineering" is really just good process documentation with AI on the receiving end. Teams that already ran tight ops were set up for this without knowing it.