In the grand scheme, this leap is evolutionary rather than revolutionary. Sure, more tokens, more context, but the challenges lie in how those tokens are processed and understood. The *interpretation* of context is where fundamental innovation will occur. Curious to see empirical studies about actual performance improvement in various applications.