Then why don’t they just do that? I read this constantly, that Google has different goals (world models, ad rev, recursive looping, quantum computing AI, magic, whatever weekly hype words go around, etc.), but they continue to try competing with model releases focused on coding performance, benchmark results and agentic harnesses.

If Googles goal was something other than those, why expend Billions in time and effort? Alternatively, they want a part of the regular model use pie, they just struggle to compete. QED, there is no secrete alternate goal here.