This is certainly a great achievement, and no one is denying that Matt’s Japanese is excellent, but this is more a matter of pronunciation and accent mastery. It’s a sign that he’s really dedicated to Japanese and likely very good, but it doesn’t prove exactly how well he knows the language. @Ryosuke didn’t mention whether or not the people he’s met have accented speech or blend in very well with co-workers and the like. However, the fact that they hold positions that involve responsibility highlights how much Japanese they need to know (they need to be capable of handling technical jargon in the workplace), and how naturally they can interact with others (because it’s not rare, for instance, to keep someone with poor knowledge of Japanese away from jobs where they might offend/confuse people with their mistakes, like in client-facing jobs). More importantly – and this is not meant as a slight to native speakers – it’s a fact that in some languages, you don’t need to be extremely proficient in order to pass as a native, because native usage within certain fields, especially in the context of social interaction, may only require relatively limited vocabulary or contain lots of mistakes that are common among native speakers.
By your logic, I must be among the best foreign French speakers in the world, because I’ve been told on multiple occasions that I have no accent (I live and study in Paris). Sure, I’m above average: I’m in a university-level science and engineering course, and my classmates tell me I speak better French than them. A French lady in her 70s told me my French is better than that of many French people. Even so, I know I have limits, especially when it comes to French philosophy and literature, which can be extremely difficult to handle because sentences can get really long and complicated, with lots of very abstract concepts and definitions. I get confused too, and that’s despite frequently outscoring much of my class in French oral assessments.
Just ‘passing for native’ doesn’t mean you’re necessarily among the absolute best: Michael Edwards is British and you can kinda hear his British accent when he speaks French if you listen closely, but there’s no doubt in my mind that his French is miles better than mine, except perhaps in the scientific domain, since he’s a literary scholar, not a scientist (but even then, I think chances that I know more are really slim). He’s part of the French Academy, who are forty of the best writers using French as their medium around the world. Bret Mayer (aka Bu-sensei) features on NHK because he passed Kanken Level 1, which requires kanji knowledge and kanji usage knowledge, including non-standard usage and probably plenty of archaic grammar too, since classical grammar is taught in Japanese schools. It’s rare even among Japanese people to pass Kanken Level 1, and the kanji you’ll find on that test are things you can only learn about in Japanese or perhaps Chinese. Almost no way you’ll find any information on them in English, Bret Mayer’s site excepted. That just proves exactly how much Japanese he needed to know to reach that level. There’s always someone better, and even if you reach the pinnacle, there’s always more to learn, because the languages we’re talking about are centuries old at the very least. 上には上がある: there’s always a higher level.
Again, I’m not saying Matt’s method doesn’t work: it worked for him, and perhaps it will work for you. Many of us agree at the very least that he’s right about immersion being very, very helpful, even if we might not agree on when it should start and whether or not other methods of study should be put aside in favour of immersion. I’m also not saying that Matt’s Japanese ‘isn’t that great’: it’s definitely better than mine at the moment, at any rate, and it’s much better than that of many foreign learners. What I am saying, however, is that you can’t know based on the experiences that he has shared that he is among the very best (which is what ‘top 3%’ would indicate), and you can’t know how he compares to native speakers in more complex/technical domains. You also can’t know if his method is truly the most effective for people in general: it’s not clear if it lines up with research – for example, more recent research has apparently shown that some sort of output is helpful for learning, so that’s already a flaw in his logic. (There’s a reference somewhere on these forums, probably under an ‘input hypothesis’ thread. I’m not the one who posted it, so I don’t remember what it was.) Also, we all know different people are comfortable with different things: I have a particular way of handling kanji, and I tend to link each and every kanji to everything else I know about its components and origins. When I tried to use that to teach my friend, he told me to let him do one thing at a time and take it slow. We’re all different, and so we need to see what suits each of us best.