Sorry, I think that's delusional and if anything maybe only applicable to the very narrow and simple problem space. Nobody writes out the "spec" for the next best inference engine or the next best distributed database.
The senior needs to figure out that the next best inference engine or the next best distributed database are what needs to be built in the first place (or beating about the bush as you said). That’s hard. Lots of conflict to be explored.
Once it’s been captured what needs to be done (spec) enabling the rest of the team (execute) is comparatively easy.
Fail to spec, and the team can’t execute. They don’t know what needs to be built.
Once it’s been captured what needs to be done (spec) enabling the rest of the team (execute) is comparatively easy.
Fail to spec, and the team can’t execute. They don’t know what needs to be built.